Network storage game allocation based on artificial intelligence
A system using AI to monitor and predict video game demand optimally distributes games across data center rack assemblies, addressing storage limitations and ensuring immediate user access by preloading popular titles, thus enhancing efficiency and reducing wait times.
Patent Information
- Application Number
- JP2024534247
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-12-08
- Filing Date
- 2022-11-23
- Publication Date
- 2024-12-05
- Estimated Expiration
- 2042-11-23
AI Technical Summary
The challenge of efficiently managing video game storage in data centers due to increasing game catalog sizes and demand variability, where individual storage units are limited and costly, leading to burdens on storage management.
Implementing a system that monitors video game usage, predicts future demand using artificial intelligence, and optimally distributes games across rack assemblies based on popularity trends, ensuring immediate access by preloading popular games on compute nodes.
Ensures real-time access to popular games by preloading them on compute nodes, optimizing storage usage, and reducing wait times for users while efficiently managing resource allocation in data centers.
Smart Images

Figure 2024545124000001_ABST
Abstract
Description
[Technical field]
[0001] The present disclosure relates to allocation of video games across data centers based on trends or trends highlighting demand and / or popularity of video games and using artificial intelligence to identify trends. [Background technology]
[0002] In recent years, online services that allow online or cloud gaming in a streaming format between cloud gaming servers and clients connected over a network have been continuously promoted. The streaming format is becoming more and more popular due to the availability of game titles on demand, the ability to execute more complex games, network connectivity between players in case of multiplayer gaming, the ability to share assets between players, the ability to share instantaneous experiences between players and / or spectators, the ability of friends to watch friends play video games, and the ability of friends to join in the ongoing gameplay of friends.
[0003] A data center may be configured with a number of computing resources to support online or cloud gaming. For example, each of the computing resources may be configured to run a video game for gameplay of a gaming application, which may then be streamed to a user. Demand for computing resources may vary depending on one or more parameters, including demand for a particular video game, the duration of the demand, the geographic area of the demand, the type of game being sought, etc.
[0004] A data center may be configured with storage that allows video games to be accessible for execution on computing resources to meet user demand for video games. The storage is typically loaded with video games from a catalog of video games supported by the data center. However, each new generation of video games requires more and more storage as the video games provide improved graphics and more detailed game environments. Video games also require faster and faster storage. It may not be practical to store the entire video game catalog in an individual storage unit because the individual storage units are limited in size due to the cost of storage. For example, each storage unit may be capable of storing about 10 percent of the video game catalog, but less as the number of video games in the catalog increases over time. This may place a strain on the storage in terms of managing the storage of video games to meet user demands, especially as the video game catalog is large and continues to grow.
[0005] It is against this background that embodiments of the present disclosure have been made. Summary of the Invention
[0006] Embodiments of the present disclosure relate to efficiently allocating video games across game cloud system data centers or the game cloud system by monitoring video game usage, system status of game cloud system computing nodes, predicting future video game usage based on historical patterns of usage, and using artificial intelligence to analyze data captured from the game cloud system and network data (e.g., scraping data from social media, websites, various forms of media, etc.) to discover and / or predict popularity trends of existing or new video games for use.
[0007] An embodiment of the present disclosure discloses a method for game allocation in a game cloud system. The method includes tracking a state of a plurality of computing nodes in a data center, the plurality of computing nodes being distributed across a plurality of rack assemblies. The method includes tracking a plurality of popularity ratings of a plurality of video games. The method includes determining a distribution of one or more video games from the plurality of video games across the plurality of rack assemblies based on the plurality of popularity ratings of the plurality of video games. The method includes storing one or more video games from the plurality of video games across the plurality of rack assemblies based on the determined distribution of the one or more video games.
[0008] Another embodiment of the present disclosure discloses a non-transitory computer readable medium storing a computer program for performing a method. The non-transitory computer readable medium includes program instructions for tracking a state of a plurality of computing nodes of a data center, the plurality of computing nodes being distributed across a plurality of rack assemblies. The non-transitory computer readable medium includes program instructions for tracking a plurality of popularity ratings of a plurality of video games. The non-transitory computer readable medium includes program instructions for determining a distribution of one or more video games from the plurality of video games across the plurality of rack assemblies based on the plurality of popularity ratings of the plurality of video games. The non-transitory computer readable medium includes program instructions for storing one or more video games from the plurality of video games across the plurality of rack assemblies based on the determined distribution of the one or more video games.
[0009] Yet another embodiment of the present disclosure discloses a computer system including a processor and a memory coupled to the processor and having instructions stored thereon, the instructions, when executed by the computer system, causing the computer system to perform a method. The method includes tracking a state of a plurality of computing nodes in a data center, the plurality of computing nodes being distributed across a plurality of rack assemblies. The method includes tracking a plurality of popularity ratings of a plurality of video games. The method includes determining a distribution of one or more video games from the plurality of video games across the plurality of rack assemblies based on the plurality of popularity ratings of the plurality of video games. The method includes storing one or more video games from the plurality of video games across the plurality of rack assemblies based on the determined distribution of the one or more video games.
[0010] Other aspects of the present disclosure will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, illustrated by way of example of the principles of the disclosure.
[0011] The present disclosure may be best understood by referring to the accompanying drawings in conjunction with the following description. [Brief description of the drawings]
[0012] [Figure 1] FIG. 1 is a diagram of a system for providing games over a network among one or more computing nodes located in one or more data centers, according to one embodiment of the present disclosure, further configured to efficiently allocate video games across the data centers to meet current and future demands of users to play current and new video games. [Figure 2A] FIG. 1 illustrates a diagram of multiple rack assemblies including multiple computing nodes in a representative data center of a game cloud system, according to one embodiment of the present disclosure. [Figure 2B]FIG. 1 is a diagram of multiple rack assemblies including multiple computing nodes in a representative data center of a game cloud system, each network storage accessible by a corresponding array of computing nodes, according to one embodiment of the present disclosure. [Figure 2C] FIG. 1 illustrates a diagram of multiple rack assemblies including multiple computing nodes configured for processing and storage, according to one embodiment of the present disclosure. [Diagram 3] FIG. 1 illustrates a system configured to implement an artificial intelligence model configured to allocate video games across a game cloud system to meet the current and future demands of users playing current and new video games, according to one embodiment of the present disclosure. [Figure 4] FIG. 1 is a flow diagram illustrating steps in a method for allocating video games across a game cloud system to meet the current and future demands of users to play current and new video games, according to one embodiment of the present disclosure. [Diagram 5] 1 illustrates components of an exemplary device that can be used to implement aspects of various embodiments of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] Although the following detailed description contains many specific details for purposes of explanation, one of ordinary skill in the art will recognize that many variations and modifications to the following details are within the scope of the disclosure. Accordingly, the aspects of the disclosure described below are set forth without any loss of generality to, and without imposing limitations on, the claims that follow this description.
[0014] Generally speaking, over time, the size of video games and game catalogs increases. Additionally, video games require faster storage. However, having faster storage in a data center is cost sensitive (i.e., expensive), so to meet the high-speed storage requirements, not all video games in the game catalog are stored on all network storage servers in the data center. An embodiment of the present disclosure provides for efficient allocation of video games across data centers (e.g., network storage) of the game cloud system by monitoring video game usage, system status of the computing nodes of the game cloud system, predicting future video game usage based on historical patterns of usage, and using artificial intelligence to analyze data captured from the game cloud system and network data (e.g., scraping data from social media, websites, various forms of media, etc.) to discover and / or predict popularity trends of existing or new video games for use. In particular, the cloud management controller is configured to keep an inventory of all available cloud gaming servers, monitor the cloud gaming servers (e.g., computing nodes), and track the status (e.g., in use or not in use) of the cloud gaming servers. Additionally, the cloud management controller is configured to monitor and track the video games stored on one or more network storage servers and / or cloud gaming servers (e.g., computing nodes) in each rack assembly. Additionally, in an embodiment, an artificial intelligence (AI) layer in the cloud management controller is configured to track the popularity of the video game titles and further configured to schedule an optimal distribution of the video games across the rack assemblies of one or more data centers of the game cloud system. For example, to meet current and projected user demand for the video game titles, each of the rack assemblies may be loaded with the most popular video game titles, less popular video game titles may be loaded into perhaps one rack assembly, other video game titles may be loaded into a selected number of rack assemblies, and still other video game titles may remain in remote storage with the corresponding video game title only being loaded into the rack assembly when a user requests that video game title. In other embodiments, when the AI layer tracks the popularity of video game titles, the AI layer is further configured to ensure that if an obscure video game title suddenly becomes trendy (i.e., becomes increasingly popular in a very short period of time), additional rack assemblies are deployed and the data center loads the video game title to meet the sudden surge in demand by users playing that video game title. In some embodiments, measuring the state of a video game when it becomes popular can be based on cloud gaming sessions, but can also be measured by external indicators such as activity on social media sites (e.g., Twitter®, Facebook®, etc.), digital or non-digital news articles, communication platforms (e.g., text messaging, voice, etc.), or past experience (e.g., a new video game within the same franchise is expected to sell as well, if not more, than a previously released video game within the franchise), digital store sales (e.g., video game sales, etc.) across one or more platforms (e.g., gaming networks, online stores, e-commerce platforms, etc.), other e-commerce data (e.g., consumer visits, etc.), etc.
[0015] Based on the foregoing general understanding of the various embodiments, illustrative details of the embodiments will now be described with reference to the various figures.
[0016] Throughout this specification, references to an "application" or a "game" or a "video game" or a "game application" or a "game title" are meant to represent any type of interactive application that is directed through the execution of input commands. By way of example only, interactive applications include applications for gaming, word processing, video processing, video game processing, and the like. Additionally, the terms introduced above are interchangeable.
[0017] 1 is a diagram of a system 100 for providing games over a network 150 among one or more computing nodes located in one or more data centers, according to one embodiment of the disclosure. The system is configured to provide games over a network among one or more cloud gaming servers, and more specifically, to efficiently allocate video games across one or more data centers to meet the current and future demand of users to play current and new video games, such that video games in a game catalog are easily and instantly accessible to users without waiting for the game cloud system to catch up with demand (e.g., without waiting for the video games to be loaded onto network storage and computing nodes). Cloud gaming involves running video games on a server to generate game-rendered video frames, which are then sent to clients for display.
[0018] It is also understood that cloud gaming and / or other services may be executed using physical machines (e.g., central processing units--CPUs--and graphics processing units--GPUs), or virtual machines, or a combination of both, in various embodiments (e.g., within a cloud gaming environment or a standalone system). For example, virtual machines (e.g., instances) may be created using a hypervisor on host hardware (e.g., located in a data center) that utilizes one or more components of a hardware layer, such as multiple CPUs, memory modules, GPUs, network interfaces, communication components, etc. These physical resources may be arranged in racks, such as a rack of CPUs, a rack of GPUs, a rack of memory, etc., and the physical resources of the racks may be accessed using a top-of-rack switch that facilitates a fabric for assembling and accessing the components used for the instances (e.g., when building the virtualized components of the instances). In general, a hypervisor can present multiple guest operating systems in multiple instances configured by virtual resources. That is, each operating system can be configured with a corresponding set of virtualized resources supported by one or more hardware resources (e.g., located in a corresponding data center). For example, each operating system may be supported by a virtual CPU, multiple virtual GPUs, virtual memory, virtualized communication components, etc. Furthermore, the configuration of the instances can be transferred from one data center to another data center to reduce latency. The instantaneous usage defined for a user or game can be used when saving a game session of the user. The instantaneous usage can include any number of configurations described herein to optimize fast rendering of video frames for a game session. In one embodiment, the instantaneous usage defined for a game or user can be transferred between data centers as configurable settings. The ability to transfer instantaneous usage settings allows efficient migration of game play from data center to data center when users connect to play games from different geographic locations.
[0019] System 100 includes a game cloud system 200 implemented across one or more data centers (e.g., data centers 1 through N). As shown, an instance of game cloud system 200 may be located in data center N providing management functions, which may be distributed across multiple instances of game cloud system 200 at each data center. In some implementations, the game cloud system management functions may be located outside of any of the data centers. For example, the game cloud system management functions may be performed by a cloud management controller, as further described in FIGS. 2A-2B.
[0020] The game cloud system 200 includes an assigner 191 configured to assign each of the client devices (e.g., 1-N) to a corresponding resource in a corresponding data center. In particular, when the client device 110 logs into the game cloud system 200, the client device 110 may be connected with an instance of the game cloud system 109 at data center N, which may be geographically closest to the client device 110. The assigner 191 may perform diagnostic tests to determine available transmit and receive bandwidth to the client device 110. The diagnostic tests may also include determining latency and / or round trip time between the corresponding data center and the client device 110. Based on the tests, the assigner 191 may assign resources to the client device 110 in a highly specific manner. For example, the assigner 191 can assign a particular data center to the client device 110. Furthermore, the assigner 191 can assign a particular computing node of a particular computing thread, of a particular streaming array, of a particular rack assembly, to the client device 110. The assignment can be performed based on knowledge of assets (e.g., games) available at the computing nodes. Previously, client devices were generally assigned to data centers and not further assigned to rack assemblies. In this way, the assigner 191 can assign a client device requesting to execute a particular computing-intensive gaming application to a computing node that may not be executing the computing-intensive application. Furthermore, load management of the assignment of the computing-intensive gaming application requested by the client can be performed by the assigner 191. For example, the same computing-intensive gaming application requested for a short period of time may be distributed to different computing nodes in one rack assembly or different rack assemblies to reduce the load on the particular computing node, computing thread and / or rack assembly.
[0021] In some embodiments, the allocation may be performed based on machine learning. In particular, resource demands may be predicted for a particular data center and its corresponding resources. For example, if it is predicted that a data center will soon handle many clients running computationally intensive gaming applications, the assigner 191 may assign client devices 110 with that knowledge and allocate resources that may not currently utilize the full capacity of that resource. In another case, the assigner 191 may switch client devices 110 from the game cloud system 200 in data center N to resources available in data center 3 in anticipation of increased load at data center N. Additionally, future clients may be assigned resources in a distributed manner such that resource load and demand may be distributed across the entire game cloud system, across multiple data centers, across multiple rack assemblies, across multiple computation threads, and / or across multiple computation nodes. For example, the client device 110 may be assigned resources from the game cloud systems in both data center N (e.g., via path 1) and data center 3 (e.g., via path 2).
[0022] Once a client device 110 has been assigned to a particular computational node of a corresponding computational thread of a corresponding streaming array, the client device 110 connects to the corresponding data center via a network, i.e., the client device 110 may communicate with a data center different from the data center that performs the assignment, such as data center 3.
[0023] Additionally, game cloud system 200 includes game allocator 300, which is configured to allocate and / or distribute the video games of the game catalog, and the computing resources necessary to execute game sessions of those video games, across data centers of the game cloud system to support current and projected (e.g., future) demand for the video games. For example, game allocator 300 may be implemented in a cloud management controller configured to manage resources in the game cloud system. In particular, the game allocator 300 may be configured to monitor video game usage and system states of compute nodes and network storage to determine current and future demand for video games and the computing resources required to meet that demand. In some embodiments, artificial intelligence (AI) may be applied to identify trends in the popularity of video games and the demand for computing resources by users due to the popularity of those video games. For example, the deep learning and / or machine learning engine 190 can be configured to implement the AI model 160, which is configured to classify and / or identify popularity trends (e.g., current and future) of one or more video game titles based on input data including metrics captured within the gaming cloud system (e.g., game sessions of one or more video games) and other data captured from network sources (e.g., social media platforms, websites, news sites, etc.), traditional sources (e.g., media outlets, etc.), other digital platforms (e.g., communication mediums (e.g., text messages, messaging, etc.)), as well as other non-digital platforms, online stores, etc. In some embodiments, the popularity trends of the video games and the predicted demand for computing resources based on the trends may be determined through rules, such as without applying artificial intelligence.
[0024] Additionally, game allocator 300 is configured to analyze results from AI model 160 in analyzer 140, including analyzing the classification and / or identified trends (e.g., indicative of the popularity of those video games) of one or more video games. Analyzer 140 is configured to determine an allocation and / or distribution of video games across one or more data centers of the game cloud system based on the classified and / or identified trends. In yet other embodiments, the allocation and / or distribution of video games across one or more data centers of the game cloud system is determined by rules applied to trends determined by artificial intelligence or through other rules.
[0025] For example, a game cloud system may have multiple data centers located around the world. One or more data centers may be located in groups, with each group serving a particular region of the world. In an embodiment, analysis of data from a first group serving a first region (e.g., identifying one or more video game trends) may be useful to a second group serving a second region, where the analysis may include allocation and / or distribution of video games across the data center(s) of the second group. That is, the allocation and prediction layers of AI may be shared among different groups of data centers of the game cloud system. By way of illustration, data analysis from the first group may indicate that video games are popular in the first region (e.g., Europe), but at the same time, the second region (e.g., North America, including the United States) is relatively inactive because people in the second region are generally asleep. Results from data analysis related to video game trends in the first group associated with the first region may be applied to the second group associated with the second region, especially if video games are also popular in the second region (e.g., North America). Specifically, the video game may be preemptively installed on a second group of rack assemblies and / or compute nodes associated with a second region in anticipation of the video game becoming popular in North America.
[0026] The system 100 provides games via the gaming cloud system 200, and according to one embodiment of the disclosure, the games are executed remotely from the client devices (e.g., thin clients) of the corresponding users playing the games. The system 100 may provide game control to one or more users playing one or more games via the cloud gaming network or gaming cloud system 200 via the network 150 in either single player or multiplayer mode. In some embodiments, the cloud gaming network or gaming cloud system 200 may include multiple virtual machines (VMs) executing on a host machine's hypervisor, where one or more virtual machines are configured to execute a game processor module that utilizes hardware resources available to the host's hypervisor. The network 150 may include one or more communication technologies. In some embodiments, the network 150 may include a fifth generation (5G) network technology having an advanced wireless communication system.
[0027] In some embodiments, communication may be facilitated using wireless technologies. Such technologies may include, for example, 5G wireless communication technologies. 5G is the fifth generation of cellular network technology. 5G networks are digital cellular networks in which the service area covered by a provider is divided into small geographical areas known as cells. Analog signals representing sound and video are digitized by the telephone, converted by an analog-to-digital converter, and transmitted as a bit stream. All 5G wireless devices within a cell communicate with local antenna arrays and low-power automatic transceivers (transmitters and receivers) within the cell by electromagnetic waves over frequency channels assigned by the transceivers from a frequency pool reused by other cells. The local antennas are connected to the telephone network and the Internet by high-bandwidth optical fiber or wireless backhaul connections. As with other cellular networks, mobile devices moving from one cell to another are automatically transferred to the new cell. Of course, 5G networks are merely an example of a type of communication network, and embodiments of the present disclosure may use previous generations of wireless or wired communication, as well as later generations of wired or wireless technologies that come after 5G.
[0028] As shown, the system 100, including the game cloud system 200, can provide access to multiple game applications. In particular, each of the client devices may be requesting access to a different game application from the cloud gaming network. For example, the game cloud system 200 can provide one or more game servers, which may be configured as one or more virtual machines running on one or more hosts to execute corresponding game applications. For example, the game server may manage virtual machines supporting game processors that instantiate instances of users' game applications. Thus, the game processors of the one or more game servers associated with the virtual machines are configured to execute multiple instances of one or more game applications associated with gameplay of multiple users. In this manner, the backend server support provides streaming of gameplay media (e.g., video, audio, etc.) for multiple gaming applications to multiple corresponding users. That is, the game servers of the game cloud system 200 are configured to stream data (e.g., rendered images and / or frames of corresponding gameplay) back to corresponding client devices over the network 150. In that way, computationally complex game applications can continue to run on the backend servers in response to controller inputs received and forwarded by the client devices. Each server can render images and / or frames, which are then encoded (e.g., compressed) and streamed to corresponding client devices for display.
[0029] In an embodiment, each virtual machine defines a resource environment capable of supporting an operating system on which gaming applications can be executed. In one embodiment, the virtual machine can be configured to emulate the hardware resource environment of a gaming console, and an operating system associated with the gaming console runs on the virtual machine to support the execution of gaming titles developed for that gaming console. In another embodiment, the operating system can be configured to emulate the native operating system environment of the gaming console, but the underlying virtual machine may or may not be configured to emulate the hardware of the gaming console. In another embodiment, an emulator application runs on top of the virtual machine's operating system, and the emulator is configured to emulate the native operating system environment of the gaming console to support gaming applications and / or video games designed for that gaming console. It should be appreciated that a variety of current and traditional gaming consoles can be emulated in the cloud-based gaming system. In this manner, users can access gaming titles from a variety of gaming consoles via the cloud gaming system.
[0030] In one embodiment, the cloud gaming network of the gaming cloud system 200 is a distributed gaming server system and / or architecture. Specifically, a distributed gaming engine that executes gaming logic is configured as a corresponding instance of a corresponding gaming application. In general, a distributed gaming engine takes each function of the gaming engine and distributes them to be executed by multiple processing entities. Individual functions may be further distributed across one or more processing entities. The processing entities may be configured in different configurations (e.g., physical hardware) and / or as virtual components or virtual machines and / or as virtual containers. A container is different from a virtual machine because it virtualizes an instance of a gaming application running on a virtualized operating system. The processing entities may utilize and / or rely on servers and their underlying hardware on one or more servers (computing nodes) of the cloud gaming network or game cloud system 200, which may be located on one or more racks. Coordination, allocation, and management of the execution of their functions for the various processing entities is performed by a distributed synchronization layer. In this manner, the distributed synchronization layer controls the execution of these functions to enable the generation of media (e.g., video frames, audio, etc.) for the gaming application in response to controller inputs by the player. The distributed synchronization layer allows critical game engine components / functions to be efficiently executed across the distributed processing entities (e.g., via load balancing) such that they are distributed and restructured for more efficient processing.
[0031] 2A is a diagram of rack assemblies 210 including compute nodes in a representative data center 200A of a gaming cloud system, according to one embodiment of the disclosure. The data centers may be distributed around the world, for example, in North America, Europe, and Japan. FIG. 2A is presented to illustrate an example configuration of a corresponding rack assembly including network storage accessible by compute nodes located in the rack assembly.
[0032] Data center 200A includes multiple rack assemblies 220 (e.g., rack assemblies 220A-220N). Each of the rack assemblies includes a corresponding network storage and multiple computational threads. For example, representative rack assembly 220N includes network storage 211 and multiple computational threads 230 (e.g., threads 230A-230N), as well as a rack controller 250 configured for internal and external network configuration of the components of rack assembly 220N. Other rack assemblies may be similarly configured with or without modification. In particular, each of the computational threads includes one or more computational nodes that provide hardware resources (e.g., processors, CPUs, GPUs, etc.). For example, computational thread 230N in the multiple computational threads 230 of rack assembly 220N is shown to include four computational nodes, although it is understood that a rack assembly may include one or more computational nodes. Each rack assembly is coupled to a cluster switch configured to provide communication with a management server configured for management of the corresponding data center. For example, rack assembly 220N is coupled to cluster switch 240N. The cluster switch also provides communication to external communication networks (such as the Internet).
[0033] In particular, the cluster fabric (e.g., cluster switches, etc.) provides communication between one or more cluster rack assemblies, distributed storage 270, and a communication network. In addition, the cluster fabric / switch also provides data center support services such as management, logging, monitoring, event generation, tracing, etc. The cluster fabric / switch provides communication to an external communication network via a router system and a communication network (e.g., the Internet). The cluster fabric / switch also provides communication to storage 270. A cluster of rack assemblies in a representative data center of a gaming cloud system may include one or more rack assemblies as a design choice. In one embodiment, the cluster includes 50 rack assemblies. In other embodiments, the cluster may include more or less than 50 rack assemblies.
[0034] In one network configuration, each rack assembly provides high-speed access to a corresponding network storage, such as within the rack assembly. In one embodiment, the high-speed access is provided via a PCIe fabric that provides direct access between the compute nodes and the corresponding network storage. In other embodiments, the high-speed access is provided via other network fabric topologies and / or networking protocols, including Ethernet, InfiniBand, Remote Direct Memory Access (RDMA) over Converged Ethernet (RoCE), and the like. For example, in rack assembly 220N, the high-speed access is configured to provide a data path 201 between a particular compute node of a corresponding compute thread to the corresponding network storage (e.g., storage 211). In particular, the PCIe fabric can provide in excess of 4 gigabytes per second (GB / s) network storage bandwidth (e.g., access) per compute node (e.g., of the rack assembly) at non-volatile memory express (NVMe) latency. Additionally, a control path 202 is configured to communicate control and / or management information between the network storage 210 and each of the compute nodes.
[0035] The streaming rack assemblies are centered around the compute nodes, which execute the gaming application, the video game, and / or stream the audio / video of the game session to one or more clients. Additionally, within each rack assembly, the video game content can be stored on storage servers that provide network storage. The network storage is equipped with a large amount of storage and a high-speed network to serve many compute nodes. Specifically, a storage protocol (e.g., a network file system, etc.) is implemented over the network fabric topology that is used to access the network storage. Data can be stored in the network storage using file storage, block storage, or object storage techniques, based in part on the underlying storage protocol implemented. For example, a PCIe fabric storage protocol may access block storage data from a network storage. The network storage may be configured for fast access, but may not be large enough to store a complete catalog of video game titles. Thus, the allocation of video games to the network storage is performed based on the current demand for video games and the predicted further demand. That is, the video games are stored in the network storage of the corresponding rack assembly to support the demand for video games by users of the game cloud system. Thus, a computing node in the rack assembly can process a game session for the video game, and the video game is stored in the network storage for immediate access and then subsequently loaded onto the computing node for execution. Depending on the license agreement, the video game may be loaded once into the network storage for the rack assembly, and subsequent copies may be loaded onto one or more computing nodes under a multi-user single copy license, or multiple instances of the video game may be loaded into the network storage rack assembly, which may then be efficiently transferred one-to-one to one or more computing nodes under a single-user single copy license.
[0036] The distributed storage 260 provides centralized storage of user data, game data, and game packages accessible by each computing node of the cluster of rack assemblies 255. The distributed storage 260 is scalable such that servers can be added or removed as required, and the storage 260 can redistribute and / or reconfigure itself. Additionally, the distributed storage is configured to provide load balancing, such that data is stored in small chunks across multiple storage servers. Management of the distributed storage can be performed from a user interface. In particular, the distributed storage 260 may be configured to store a complete complement of video game titles (e.g., game applications) for the game cloud system 200 (e.g., approximately 200 terabytes [TB] of storage to store a complete catalog of game applications). The distributed storage 260 may also be configured to store user and game data for users of the game cloud system 200. The user and game data may include user save data, corresponding game application pause / resume data, download data, and so forth (images, photos, etc.). To illustrate, each user may require approximately 50 gigabytes (GB) of storage per user. In one embodiment, the distributed storage 260 stores pause / resume information associated with corresponding game titles (e.g., 8 GB per game title).
[0037] As shown, the cloud management controller 210 of the data center 200 communicates with the assigner 191 (shown in FIG. 1 ) to allocate resources to the client devices 110. In particular, the cloud management controller 210 may coordinate with an instance of the gaming cloud system 200′ and with the initial instance of the gaming cloud system 200 (e.g., of FIG. 1 ) to allocate resources to the client devices 110. In an embodiment, the allocation is performed based on asset awareness, such as knowing what resources and bandwidth are needed and that exist in the data center. Thus, an embodiment of the present disclosure is illustratively configured to assign the client device 110A to a particular compute node 232A of a corresponding compute thread 231A of a corresponding rack assembly 220B.
[0038] The cloud management controller 210 is also configured to collect and analyze data (e.g., internal data related to the gaming cloud system, external data such as network data, etc.) to identify trends in popularity of video games for the purpose of allocating video games across rack assemblies (e.g., network storage, compute nodes, etc.) of one or more data centers of the gaming cloud system to meet current and predicted (e.g., future) demand for playing those video games (e.g., game sessions, etc.). In particular, the cloud management controller 210 is configured to track and manage an inventory of all available cloud gaming servers (e.g., compute nodes), monitor the cloud gaming servers, and track the status (e.g., in use or not in use) of the cloud gaming servers. Additionally, the cloud management controller is configured to monitor and track video games stored on one or more network storage servers and / or compute nodes in each rack assembly. Additionally, the cloud management controller 210 may be configured to track game sessions implemented across the cloud gaming servers. In one embodiment, a game allocator 300 may be included within or in communication with the cloud management controller 210 to perform operations necessary to allocate video games to the network storage and / or compute nodes of the rack assemblies of one or more data centers of the gaming cloud system, as further described in FIG. 3 below. In some embodiments, artificial intelligence is included within the game allocator 300 to perform the operations necessary to allocate video games to network storage. In particular, the AI layer within the cloud management controller 210 and / or the game allocator 300 is configured to track the popularity of video game titles and further configured to schedule an optimal distribution of the video games across the rack assemblies of one or more data centers of the game cloud system. That is, the AI is used to determine the popularity of the video games and to determine the number of rack assemblies for storing the video games based on the determined popularity.
[0039] 2B is a diagram of multiple rack assemblies 221 including multiple computing nodes in an exemplary data center 200B of a gaming cloud system, where each network storage is accessible by a corresponding array of computing nodes, according to one embodiment of the disclosure. Data center 200B is similar to data center 200A, where like-numbered components have similar functions. However, data center 200B has a different configuration of rack assemblies than the rack assemblies of data center 200A, where the network storage is accessed by a single streaming array of computing nodes, as described below.
[0040] The data center 200 includes multiple rack assemblies 221 (e.g., rack assemblies 221A to 221N). Each of the rack assemblies includes one or more streaming arrays, each including a corresponding network storage and multiple computational threads. For example, the representative rack assembly 221N includes streaming arrays 225A to 225N. In one embodiment, the rack assembly 221N includes two streaming arrays, each including a network storage and multiple computational threads. For example, the streaming array 225N includes multiple computational threads 235 that access the network storage 211-N. In particular, each of the computational threads includes one or more computational nodes that provide hardware resources (e.g., processors, CPUs, GPUs, etc.). For example, although the computational thread 235X of the streaming array 225N is shown to include four computational nodes, it is understood that the rack assembly can include one or more computational nodes.
[0041] Each rack assembly is coupled to a cluster switch configured to provide communication with a cloud management controller 210 configured for management of a corresponding data center, as described above. For example, rack assembly 221N is coupled to cluster switch 240N. The cluster switch also provides communication to other rack assemblies (e.g., via a corresponding cluster switch) and to an external communication network (e.g., the Internet, etc.).
[0042] In one network configuration, each streaming array of a corresponding rack assembly provides high-speed access to a corresponding network storage, as described above. This high-speed access is provided via a PCIe fabric that provides direct access between the computing nodes and the corresponding network storage. The computing nodes can execute game applications and stream audio / video of game sessions to one or more clients, and the corresponding network storage (e.g., a storage server) holds the game applications, game data, and user data. For example, in the streaming array 225A of the rack assembly 221N, the high-speed access is configured to provide a data and control path 201A between a particular computing node of a corresponding computing thread and the corresponding network storage (e.g., storage 211A). The path 201N is also configured to communicate control and / or management information between the network storage 211N and each of the computing nodes of the streaming array 225N.
[0043] As previously discussed, the cloud management controller 210 of the data center 200B communicates with the assigner 191 to allocate resources to the client devices 110 supporting the gaming cloud system 200′ and / or 200. In an embodiment, the allocation is performed based on asset awareness, such as knowing what resources and bandwidth are needed and present in the data center. Thus, for illustrative purposes, an embodiment of the present disclosure is configured to assign the client device 110C to a particular compute node 232B of a corresponding compute thread 231B of a corresponding streaming array of the rack assembly 221B.
[0044] The cloud management controller 210 and / or the game allocator 300 are also configured to collect and analyze data (e.g., internal data related to the game cloud system, external data such as network data, etc.) to identify trends in popularity of video games for the purpose of allocating video games across rack assemblies (e.g., network storage, compute nodes, etc.) of one or more data centers of the game cloud system to meet current and predicted demand for playing those video games (e.g., game sessions). In some embodiments, artificial intelligence is implemented to analyze the data to identify trends in popularity of video games for the purpose of determining an appropriate allocation of video games to the network storage of the rack assemblies of one or more data centers of the game cloud system. In particular, the AI layer within the cloud management controller 210 and / or the game allocator 300 is configured to track the popularity of video game titles and further configured to schedule an optimal distribution of video games across rack assemblies of one or more data centers of the game cloud system. That is, the AI is used to determine the popularity of the video games and to determine the number of rack assemblies for storing the video games based on the determined popularity.
[0045] 2C is a diagram of a data center 200C including multiple rack assemblies including multiple compute nodes configured for processing and storage, according to one embodiment of the disclosure. Data center 200C is similar to data centers 200A and 200B, where like-numbered components have similar functions. However, data center 200C has a different configuration of rack assemblies, whereby the rack assemblies are configured with standalone compute nodes, each of which includes dedicated storage (i.e., does not rely on network storage).
[0046] In particular, data center 200C includes multiple rack assemblies 222 (e.g., rack assemblies 222A through 222N). Each of the rack assemblies includes multiple computational threads 237, each of which includes one or more computational nodes that provide hardware resources. For example, multiple computational threads 237X of rack assembly 222N are shown to include four computational nodes (including computational node 233), but it is understood that a rack assembly may include one or more computational nodes. The computational nodes may execute gaming applications and stream audio / video of gaming sessions to one or more clients. More specifically, each computational node may be configured as a standalone system (e.g., a game console, etc.) including a central processing unit (CPU), a graphics processing unit (GPU), volatile memory including random access memory (RAM) used during operation of the computational node (e.g., for quickly accessing frequently accessed data, programs, processes, etc.), non-volatile memory including local or system memory, etc. For purposes of illustration, a compute node 233 of a compute thread 237X of a rack assembly 222N includes, in part, a CPU 271 and a system memory 272 (eg, non-volatile memory).
[0047] Each rack assembly is coupled to a cluster switch configured to provide communication with a cloud management controller 210 configured for management of a corresponding data center, as described above. For example, rack assembly 222N is coupled to cluster switch 240N. The cluster switch also provides communication to other rack assemblies (e.g., via a corresponding cluster switch) and to an external communication network (e.g., the Internet, etc.).
[0048] As previously described, the cloud management controller 210 of the data center 200C communicates with the assigner 191 to allocate resources to the clients 110 supporting the gaming cloud system 200′ and / or 200. For example, the client device 110C may be assigned to a particular computing node 232C of a corresponding computing thread 231C of the rack assembly 222B. The cloud management controller 210 and / or the game allocator 300 may also be configured to collect and analyze data (e.g., internal and external data) to identify trends in popularity of video games and meet current and projected demand for playing those video games across the rack assemblies (e.g., network storage, computing nodes, etc.) of the data center(s) of the gaming cloud system. Artificial intelligence (e.g., an AI layer within the cloud management controller or the game allocator) may be implemented to analyze the data to identify trends in games for purposes of determining appropriate allocation of video games to the network storage and / or computing nodes of the rack assemblies of the data center(s) of the gaming cloud system. That is, AI is used to determine the popularity of a video game and, based on the determined popularity, determine the number of rack assemblies and / or computing nodes to store the video game.
[0049] FIG. 3 is a diagram of a system configured to implement an artificial intelligence model configured to allocate video games across a game cloud system to meet current and future demands of users to play current and new video games, according to one embodiment of the disclosure. In an embodiment of the disclosure, trends identifying popularity of video games can be classified using AI model 160 and analyzed using analyzer 140 to allow allocation of video games to rack assemblies of a game cloud system in which one or more data centers may be located. For illustrative purposes, FIG. 3 is described within the context of game allocator 300, although it is understood that operations performed within the system of FIG. 3 may be performed by cloud management system 210 and / or game allocator 300 of FIGS. 1, 2A, and 2B.
[0050] In particular, the game allocator 300 is configured to provide an appropriate allocation of video games across the game cloud system's data centers (e.g., network storage and / or compute nodes) to meet users' current and projected demand for video games in the game cloud system's game catalog, and the rack assemblies may be comprised of network storage for access by the compute nodes, or may be comprised of many individual computers and storage servers that do not rely on network storage. In some embodiments, the distribution of video games across one or more data centers, and more particularly across the game cloud system's rack assemblies, is configured to enable the game cloud system to provide the most efficient use and / or optimal resource usage, ensuring the quickest (e.g., real-time) access to video games requested by users of the game cloud system.
[0051] In particular, capture engine 340 of game allocator 300 may be configured to capture game cloud network data, network data, and other data to provide input to AI model 160 for classification of video game popularity trends. That is, capture engine 340 may be configured to capture and / or receive as input any data that may be used to determine how to allocate video games across rack assemblies of the game cloud network. Selected portions of the captured data may be analyzed to determine an appropriate allocation of video games across rack assemblies of the game cloud system.
[0052] For example, the capture engine 340 is configured to capture data related to services provided by the game cloud system 200, including one or more data centers, such as providing game sessions of video games to users. The data may include metrics defining video game usage (e.g., a history of game sessions of one or more video games), system status of the compute nodes in the game cloud system, and network storage status, including storage of video games in each network storage of the rack assemblies of the data center supporting the game cloud system. Still other data may be collected that is relevant to assist in determining an appropriate allocation of video games to network storage to meet current and projected demand for video games (e.g., based on popularity trends of those video games). Additionally, the data center tracker 343 is configured to track the data captured from the game cloud system for purposes of determining an appropriate allocation of video games to meet current and projected demand for video games.
[0053] Further, the capture engine 340 is configured to collect telemetry data 315 from various sources 310a-n. As shown, a stream of telemetry data 315 is captured. For example, the telemetry data may include network data, such as social network platform data (e.g., Twitter, Facebook, etc.), social media data, website data, media data, audio data, voice data, text data, messaging data, other forms of digital data suitable for capture, and other data received over the network 150 (e.g., the Internet). For example, the network data may be captured by web scraping techniques used to automatically collect publicly available information accessed via any communications network (e.g., the Internet). As an example, the captured network data may be formatted using a text-based markup language. In some implementations, the web scraping techniques implement application programming interfaces (APIs) used to communicate with websites and / or social network platforms or any services accessible via the network. In other implementations, the capture engine 340 is configured to capture non-digital data. For example, the data may be analyzed by a machine and / or a human and converted into a digital format before being input to the capture engine 340.
[0054] The relevant features from the captured data (e.g., telemetry data 315 and data from the game cloud system 305) identified by the feature extractor 345A are provided as input to the deep learning / machine learning engine 190. The feature extractor 345A is configured to analyze the captured data and identify and / or extract features from the data, which are relevant for use in the allocation of video games to the network storage of the game cloud system. For example, the features may include parameters, time, video games, game states, game sessions, words, symbols, relevant information related to the popularity of the video games, etc. In another embodiment, the extraction of the features may be performed by the deep / machine learning engine 190. In particular, the feature extractor 345B is configured to analyze the captured data and identify and / or extract features from the data. That is, the feature extractor 345B is also capable of learning features that may be useful in the allocation of video games to the network storage of the game cloud system.
[0055] As shown, deep / machine learning engine 190 is configured for implementation (e.g., implementation phase) of classifying video game popularity trends based on an input set of data (e.g., extracted features), but may be configured for training at another phase (e.g., training phase). In one embodiment, AI model 160 is a machine learning model configured to apply machine learning to classify video game popularity trends. In another embodiment, AI learning model is a deep learning model configured to apply deep learning to classify video game popularity trends, where machine learning is a subclass of artificial intelligence and deep learning is a subclass of machine learning.
[0056] Purely for illustrative purposes, the deep / machine learning engine 190 may be configured as a neural network used to implement the AI model 160 according to one embodiment of the present disclosure. In general, a neural network represents a network of interconnected nodes that respond to inputs (e.g., extracted features) and generate outputs (e.g., a classification of trends indicative of the popularity of video games). In one implementation, the AI neural network includes a hierarchy of nodes. For example, there may be an input layer of nodes, an output layer of nodes, and an intermediate or hidden layer of nodes. The input nodes are interconnected to hidden nodes in the hidden layer, and the hidden nodes are interconnected to output nodes. The interconnections between the nodes may have numerical weights that may link multiple nodes together between inputs and outputs, such as when defining the rules of the AI model 160.
[0057] In particular, AI model 160 is configured to apply rules that define relationships between features and outputs (e.g., popularity of video games, trends in popularity of video games, etc.), where features may be defined within one or more nodes located at one or more hierarchical levels of AI model 160. Rules link features (defined by nodes) across layers of the hierarchy such that a given input set of data leads to a particular output (e.g., trend classification 350) of AI model 160 (e.g., trends indicative of current and / or predicted popularity of video games). For example, rules may link one or more features or nodes between inputs and outputs (e.g., using relationship parameters, including weights) throughout AI model 160 (e.g., at hierarchical levels) such that one or more features make up the rules that are learned through training of AI model 160. That is, each feature may be linked with one or more features of other layers, where one or more relationship parameters (e.g., weights) define the interconnections between features of other layers of AI model 160. Thus, each rule or rule set corresponds to a classified output. For example, the resulting output 350 according to the rules of the AI model 160 may predict certain trends in video games that are indicative of the current and / or projected popularity of video games.
[0058] Thus, during the implementation stage, AI model 160 is used to predict and / or classify one or more video game popularity trends based on the configuration of AI model 160. That is, for a given set of extracted features provided as input to AI model 160, a corresponding video game trend classification 350 may be output by the AI model.
[0059] Additionally, the output from the AI model 160 (e.g., trend classification 350) may be used by the analyzer 140 to determine a course of action to take for a given set of inputs (e.g., extracted features). For example, the action may include a distribution pattern for allocating video games in the game catalog to network storage in rack assemblies of one or more data centers of the game cloud system.
[0060] Specifically, the analyzer 140 may include one or more functions performed in trend classification 350 to discover, confirm, and / or predict popularity trends for existing or new video games. In one embodiment, the analyzer 140 is configured to perform trend analysis 361. The analysis may be performed to identify relevant information related to the corresponding video game, such as when the video game was released, the game title, any franchise affiliation, developer, the type of users who play the video game, user loyalty, etc., average duration of a game session, demographics, etc. This information may be useful in determining allocation of video games within the game cloud system to meet current and near future (e.g., next 2-6 hours) demand for video games.
[0061] Additionally, analyzer 140 may be configured to perform trend prediction and / or matching 363. For example, video game trend classification 350 may indicate the current popularity of the video game. Further analysis of the trend classification and / or other related data (e.g., information provided from performing trend analysis 361) may predict the continuity of popularity, or predict whether popularity will increase in the near or distant future, or predict whether popularity will decrease in the future or distant future, or predict whether popularity will remain stable in the near or distant future. For example, if an important national event (e.g., watching a sporting event, giving a speech, etc.) is scheduled for that evening, it may be predicted that the popularity of the video game and other video games may decline at least in the hours prior to the event as users abandon their gaming sessions to attend the national event (e.g., watching a sporting event, giving a speech, etc.). Additionally, the trend classifications and other data of the first video game may closely mirror the trend classifications and other data of the second video game. Thus, the predicted popularity of the first video game over time may be estimated and / or closely match the actual popularity of the second video game over time. All of this information may be useful in determining the allocation of video games within the game cloud system to meet the demand for video games in the present and near future (e.g., within the next 2-6 hours) or distant future (e.g., the next few days, or weeks, or months, etc.).
[0062] Additionally, the analyzer 140 may be configured to perform historical trend analysis 365. For example, historical trend analysis may be used to predict future video game usage based on historical patterns. In that case, the popularity of a new release of a new version of a video game may follow or closely match the actual popularity of previous versions of the video game. For example, in a highly popular video game franchise, the popularity of later releases may be equal to or even greater than the popularity of previous releases. This information may be useful in determining video game allocations across the game cloud system to meet current and near-future (e.g., within the next 2-6 hours) or distant-future (e.g., next few days, or weeks, or months) video game demand.
[0063] The analyzer 140 may also be configured to perform additional responses (e.g., actions) via the action generator 370. In particular, the analyzer 140 may be configured to perform pattern generation 367 to determine a distribution pattern for allocating video games in the game cloud system. For example, the analyzer may determine the allocation pattern based on one or more of the trend classification 350, information obtained from trend analysis 361, information obtained from trend prediction and / or matching 363, and information obtained from historical trend analysis 365.
[0064] In one embodiment, the action generator 370 includes a rack distributor 373 configured to allocate the video games to the network storage of the rack assemblies of one or more data centers of the game cloud system. That is, based on the pattern determined by the pattern generation 367, the cloud management controller may send an instruction to load the corresponding video games into the network storage of one or more corresponding rack assemblies, which may be located in one or more data centers of the game cloud system. Thus, based on a trend indicating the current and / or popularity of the corresponding video game, the distribution of the video game may be allocated to meet the current and / or predicted demand to play the video game. This video game allocation may be performed for each video game having a corresponding trend classification output from the AI model 160.
[0065] In another embodiment, the action generator 370 includes a video game node preloader 375 that is configured to load a corresponding video game onto one or more computational nodes to facilitate the allocation of the computational nodes to any new user requesting to play the video game. The preloading of the video game onto the nodes is performed based on trend classifications that indicate the current and / or predicted popularity of the video game, which may correspond to the current and / or predicted demand for playing the video game by users. By preloading the video game onto the computational nodes, a new user does not have to wait for the video game to be loaded onto the computational node assigned to the user. As video games become more complex, this loading time may become significant (e.g., 10-20 minutes) to download all packages required for the video game onto the computational node. In that way, preloading allows a user to immediately play the video game without waiting for the game to be loaded.
[0066] 1, 2A, 2B, and 3, and the various modules of the system 100, data centers 200A and 200B, and game allocator 300, flow diagram 400 discloses a method for allocating video games across a game cloud system to meet the current and future demands of users to play current and new video games, according to one embodiment of the present disclosure. In particular, flow diagram 400 can be implemented to provide an appropriate allocation of video games across the data centers (e.g., network storage) of the game cloud system based on the popularity of the video games. For example, the popularity of a video game may be determined by monitoring video game usage, system status of the computing nodes of the game cloud system, predicting current and future video game usage based on historical patterns of usage, and using artificial intelligence to analyze data captured from the game cloud system and network data (e.g., scraping data from social media, websites, various forms of media, etc.) to discover and / or predict trends in the popularity of existing or new video games. Flow diagram 400 may be implemented within a backend server (e.g., within the game allocator 300 in combination with the game cloud system 200 and / or deep and / or machine learning engine 190 of the cloud management controller 210), as described above.
[0067] At 410, the method includes tracking a state of a plurality of compute nodes in the data center, including monitoring the compute nodes. The compute nodes may be distributed across a plurality of rack assemblies, and the rack assemblies may be configured in one or more data centers of a game cloud system that provides access to video games of a video game catalog. For example, the game cloud system may be configured to run video games on the compute nodes to provide game sessions to users playing those video games.
[0068] In one embodiment, the tracking includes tracking the usage (e.g., active or inactive) of each of the plurality of computing nodes. For example, the information may include whether the corresponding computing node is active or inactive, whether a video game is loaded on the corresponding computing node, and whether a game session is actively running to support a game session with a user assigned to the corresponding computing node. Additionally, the tracking includes tracking an inventory of cloud gaming servers or computing nodes, including the number of computing nodes in the gaming cloud system, the number of active computing nodes actively supporting game sessions, and the number of available computing nodes that can be assigned to a new user requesting to play the corresponding video game. The tracking may include tracking current and / or historical (e.g., past) usage of the computing nodes.
[0069] Also, the tracking may include tracking distribution of the video games of the video game catalog across multiple network storages of rack assemblies in one or more rack assemblies of the game cloud system. Since the network storage is limited, each network storage (e.g., of the corresponding rack assembly) may not be able to store the complete catalog of video games, especially when fast access of the network storage is important (i.e., when fast storage may require increased costs of the network storage and may limit the size of the network storage). Also, not all video games of the video game catalog are loaded onto the network storage. For example, older video games may have less demand for game sessions, and at a particular time, there may be no demand and no storage servers have the video game loaded or only a minimum number of storage servers (e.g., 1-3) have the video game loaded. On the other hand, very popular video games of the video game catalog may be loaded onto each network storage in one or more data centers of the game cloud system.
[0070] At 420, the method includes tracking the popularity of a plurality of video games in a catalog of video games supported by the game cloud system. More specifically, a popularity rating may be tracked, the rating indicating a degree of popularity for the corresponding video game. In some implementations, the popularity rating of the video game may correspond to demand for game sessions by users requesting to play the video game. That is, as the popularity rating increases (i.e., indicating the popularity of the video game is improving), there is a corresponding increase in game sessions requested by users of the game cloud system. In one embodiment, the popularity rating is tracked by tracking game sessions of the corresponding video games in the game cloud system to determine current and / or predicted popularity, and the game sessions are processed by corresponding computational nodes of the game cloud system.
[0071] In other embodiments, as discussed above, popularity tracking may include collecting and / or tracking network data, including information obtained from a gaming cloud system (e.g., game sessions) and / or data captured from social networking platforms, websites, media sites, media outlets, etc. Other non-digital forms of data may also be captured and / or tracked.
[0072] In some embodiments, tracking popularity includes determining trends indicative of the current and / or projected popularity of the video game. To illustrate, based on the captured and / or tracked information (e.g., tracking one or more game sessions of the video game across computational nodes of one or more data centers of the game cloud system), a trend indicative of popularity (e.g., current popularity or projected popularity in the near or distant future) may be determined. In one embodiment, the trend may correspond to a popularity rating of the video game. For example, the trend may indicate an increase in the popularity rating of the video game. The trend may be determined based on one or more game sessions of the tracked video game, and the popularity rating may correspond to an increase in demand for game sessions for that video game. The trend may also be determined based on other captured and tracked information, as discussed above. Also, the trend may indicate a decrease in the popularity rating of the video game, which may correspond to a decrease in demand for game sessions for that video game.
[0073] In some embodiments of the present disclosure, artificial intelligence is implemented to track and / or determine the popularity of video games in the game catalog. Based on the determined popularity, further application of rules and / or AI may be implemented to determine and / or schedule an optimal distribution of video games across rack assemblies (e.g., in corresponding network storage) of one or more data centers of the game cloud system to meet current and / or projected demand for those video games.
[0074] As previously discussed, AI may be used to determine and / or track the popularity of video games in the catalog of video games supported by the game cloud system. For example, the game allocator 300 and / or the cloud management system 210 may be configured to use AI to determine and / or track the popularity of video games. These results may be used to take further action (e.g., determined through rules or additional application of AI) to schedule an optimal distribution of video games across rack assemblies of one or more data centers of the game cloud system based on the popularity of those video games. For example, a popular video game may result in an increase in the number of rack assemblies that load that video game for access by compute nodes supporting game sessions of that video game. Conversely, a less popular video game may result in a decrease in the number of rack assemblies that load that video game.
[0075] In particular, data from the game cloud system (e.g., game sessions of the video game) and / or telemetry data captured over a network (e.g., data scraped from multiple sites, social network platforms, etc. accessible over one or more networks (e.g., using an API) may be captured and analyzed to extract relevant features. Those features are input to a deep learning and / or machine learning engine that implements an AI model configured to classify a corresponding video game and a popularity trend of each video game in a catalog of video games supported by the corresponding game cloud system. That is, the AI model is configured to classify a trend of the video game based on the features (e.g., the AI model provides the trend as an output). The trend may be indicative of a current and / or predicted popularity of the video game. More specifically, the trend may be indicative of a popularity rating of the video game. For example, the trend may indicate that the popularity (and corresponding popularity rating) of the video game is increasing, whereby the popularity rating corresponds to an increase in demand for game sessions of the video game of the game cloud system. Also, the trend may indicate that the popularity (and corresponding popularity rating) of the video game is decreasing, whereby the popularity rating corresponds to a decrease in demand for game sessions of the video game of the game cloud system.
[0076] At 430, the method includes determining a distribution of one or more video games from the plurality of video games across the plurality of rack assemblies based on a plurality of popularity ratings of the plurality of video games. In particular, the allocation of the video games to the network storage of the rack assemblies in the one or more data centers of the game cloud system is performed to meet current and projected user demand for the one or more video games based on determined trends indicative of the popularity of those video games. For example, based on the determined trends (e.g., determined by AI) for a corresponding video game, the allocation may indicate an increase in the number of network storage servers and corresponding rack assemblies configured to load and / or store the video game when the trends indicate an increase in popularity, or the allocation may indicate a decrease in the number of network storage servers and corresponding rack assemblies configured to load and / or store the video game when the trends indicate a decrease in popularity.
[0077] As a result, a distributed pattern indicating the allocation of video games to the network storage of the game cloud system may dictate that each rack assembly in one or more data centers should be loaded with the most popular video games (e.g., the currently most talked about video games) and only one or two rack assemblies should be loaded with less popular video games (e.g., older video games). As previously described, the entire catalog of video games is stored in distributed storage separate from the network storage of the rack assemblies. Still other video games may be loaded into a selected and / or determined number of rack assemblies based on their popularity ratings. Thus, video games that are in demand or predicted to be in demand are loaded into the network storage, while the remaining video games are stored in distributed storage for downloading to the network storage when demand for those video games increases.
[0078] At 440, the method includes storing one or more video games from the plurality of video games in a game catalog across the plurality of rack assemblies based on the determined distribution of the one or more video games. In that way, proper allocation of the video games ensures that current and future demand for game sessions of current and / or new video games can be supported by the game cloud system. For example, proper allocation of a popular video game ensures that all users requesting that video game are assigned to a compute node that executes a game session of that video game. Also, proper allocation of a less popular video game prevents over-allocation of resources to that video game while ensuring that users requesting that video game can be assigned to a compute node that executes a game session established for that video game. In that way, by not over-provisioning computing resources (e.g., network storage) for a less popular video game, it may remain available to process other more popular video games.
[0079] In one embodiment, the classified and / or identified trends and subsequent allocation of video games are performed to ensure that if an obscure video game title suddenly becomes trendy (i.e., becomes increasingly popular in a very short period of time), additional rack assemblies are placed and the data center loads the video game title to satisfy the spike in demand by users who play the video game title. This may require de-emphasizing some video games by removing them from rack assemblies such that the number of rack assemblies supporting the corresponding de-emphasized video games is reduced. The space recovered on the network storage can be used to load video games that are predicted to become trendy or that have been identified as becoming trendy.
[0080] In another embodiment, artificial intelligence may be implemented to predict an increase in demand for a video game. For example, active game sessions of a video game may be tracked across multiple computing nodes of a game cloud system to determine trends indicative of the popularity of the video game. The trends may also predict future periods of high demand for game sessions of the video game. In that way, video games may be loaded across an increasing number of rack assemblies to support high demand by users wanting to play the video games.
[0081] In another embodiment, the popularity of a video game may be determined and / or matched to the popularity of a related video game. That is, the demand for a second video game may be inferred from the demand for a first video game. For example, the second video game may be a new version of a video game franchise. The second video game (new release) is predicted to perform as well (e.g., be as popular) as the last released video game in the franchise or the first video game based on previous popularity tracking. More specifically, a first history of a first popularity rating of a first video game is tracked across a plurality of compute nodes in a data center for at least a portion of the first video game's existence. A first trend is determined for the first video game, the first trend may indicate that the first popularity rating of the first video game has increased over a first period of time during the portion of the first video game's existence. A first number of rack assemblies may be configured to store the first video game based on the first trend, previously or newly determined, as indicated in the first history. A second trend can then be determined that indicates that a second popularity rating of the second video game over a second time period will be similar to the first trend of the first video game. For example, the two trends may have matching parameters implemented via an AI model, or the first video game and the second video game may be part of a video game franchise, and it can be inferred that subsequent releases will be similarly, if not more, popular than previous releases. In that way, an appropriate number of rack assemblies for storing the second video game (e.g., a second number of rack assemblies) can be determined based on the second trend, the second number of rack assemblies being approximately the same as the first number used to support game sessions of the first video game.
[0082] In yet another embodiment, an AI is implemented to identify trends that indicate an increasing popularity rating of the video game based on capturing and tracking game cloud data and / or network data, etc., where the trends may be determined by the AI or rule implementation. To meet user requests and provide users with real-time access to the video games (i.e., playing the video game immediately or shortly after requesting a game session of the video game), the video games may be loaded into the network storage of the rack assembly (i.e., making the video games immediately accessible to computing nodes of that rack assembly) and may be pre-loaded onto the computing nodes of the rack assembly. In that way, a user requesting a game session can be immediately assigned to a computing node on which the video game has been preloaded. During the load time, the video game must be loaded from the network storage onto the computing node, which may take several minutes or more, but the user does not have to wait for this load time. It will be appreciated that video games may be preloaded onto one or more computing nodes of a rack assembly, and can be loaded onto the network storage of multiple rack assemblies, which can then be preloaded in turn onto multiple computing nodes of multiple rack assemblies to meet the anticipated demand for video games.
[0083] FIG. 5 illustrates components of an exemplary device 500 that can be used to implement aspects of various embodiments of the present disclosure. For example, FIG. 5 illustrates an exemplary hardware system suitable for providing video game allocations across a data center of a gaming cloud system to meet the current and future demands of users playing current and new video games, according to embodiments of the present disclosure. The block diagram illustrates device 500, which may incorporate or be a personal computer, a server computer, a gaming console, a mobile device, or other digital device, each suitable for implementing embodiments of the present invention. Device 500 includes a central processing unit (CPU) 502 for executing software applications and optionally an operating system. CPU 502 may be comprised of one or more homogeneous or heterogeneous processing cores.
[0084] According to various embodiments, CPU 502 is one or more general-purpose microprocessors having one or more processing cores. Further embodiments can be implemented using one or more CPUs with microprocessor architectures specifically adapted for highly parallel, computationally intensive applications (e.g., media and interactive entertainment applications) for applications configured to perform graphics processing during game execution.
[0085] Memory 504 stores applications and data used by CPU 502 and GPU 516. Storage 506 provides non-volatile storage and other computer-readable media for applications and data, and may include fixed disk drives, removable disk drives, flash memory devices, and CD-ROM, DVD-ROM, Blu-ray, HD-DVD, or other optical storage devices, as well as signal transmission and storage media. The user input devices 508 communicate user input from one or more users to the device 500, examples of which may include a keyboard, a mouse, a joystick, a touchpad, a touch screen, a still or video recorder / camera, and / or a microphone. The network interface 509 enables the device 500 to communicate with other computer systems over an electronic communication network, which may include wired or wireless communications over local area networks and wide area networks such as the Internet. The audio processor 512 is adapted to generate analog or digital audio output from instructions and / or data provided by the CPU 502, the memory 504, and / or the storage 506. The components of the device 500, including the CPU 502, the graphics subsystem including the GPU 516, the memory 504, the data storage 506, the user input devices 508, the network interface 509, and the audio processor 512, are connected via one or more data buses 522.
[0086] A graphics subsystem 514 is further coupled to the data bus 522 and the components of the device 500. The graphics subsystem 514 includes at least one graphics processing unit (GPU) 516 and a graphics memory 518. The graphics memory 518 includes a display memory (e.g., a frame buffer) used to store pixel data for each pixel of an output image. The graphics memory 518 may be integrated into the same device as the GPU 516, may be coupled as a separate device from the GPU 516, and / or may be implemented within the memory 504. Pixel data may be provided to the graphics memory 518 directly from the CPU 502. Alternatively, CPU 502 may provide data and / or instructions defining a desired output image to GPU 516, from which GPU 516 generates pixel data for one or more output images. The data and / or instructions defining the desired output image may be stored in memory 504 and / or graphics memory 518. In an embodiment, GPU 516 includes 3D rendering capabilities to generate pixel data for output images from instructions and data defining scene geometry, lighting, shading, texture, motion, and / or camera parameters. GPU 516 may further include one or more programmable execution units capable of executing shader programs.
[0087] Graphics subsystem 514 periodically outputs pixel data for images from graphics memory 518 to be displayed on display device 510 or projected by a projection system (not shown). Display device 510 may be any device capable of displaying visual information in response to a signal from device 500, including CRT, LCD, plasma, and OLED displays. Device 500 may provide analog or digital signals, for example, to display device 510.
[0088] In other embodiments, graphics subsystem 514 includes multiple GPU devices that are combined to perform graphics processing for a single application running on a corresponding CPU. For example, multiple GPUs can perform multi-GPU rendering of geometry for an application by pre-testing geometry against a region of the screen (which may be interleaved) before rendering objects for an image frame. In another example, multiple GPUs may perform alternating frame rendering, where GPU1 renders a first frame, GPU2 renders a second frame, and so on in successive frame cycles until the last GPU is reached. The first GPU then renders the next video frame (e.g., if there are only two GPUs, then GPU1 renders the third frame). That is, the GPUs rotate as they render frames. Rendering operations may overlap, such that GPU2 may begin rendering the second frame before GPU1 has finished rendering the first frame. In another embodiment, different shader operations in the rendering and / or graphics pipeline can be assigned to multiple GPU devices. A master GPU performs main rendering and compositing. For example, in a group including three GPUs, master GPU1 can perform main rendering (e.g., first shader operations) and compositing of outputs from slave GPU2 and slave GPU3, slave GPU2 can perform second shader (e.g., fluid effects such as rivers) operations, slave GPU3 can perform third shader (e.g., particle spraying) operations, and master GPU1 composites the results from each of GPU1, GPU2, and GPU3. In that way, different GPUs can be assigned to perform different shader operations (e.g., flagging, wind, spray, fire, etc.) to render a video frame. In yet another embodiment, each of the three GPUs can be assigned to different objects and / or portions of the scene corresponding to the video frame. In the above embodiments and implementations, these operations can occur in the same frame cycle (concurrently in parallel) or in different frame cycles (sequentially in parallel).
[0089] Accordingly, the present disclosure describes methods and systems configured to allocate video games across a game cloud system to meet the current and future demands of users to play current and new video games.
[0090] It should be understood that the various embodiments defined herein may be combined or assembled into specific implementations that use various features disclosed herein. Thus, the examples provided are only some of the possible examples and are not intended to limit the various implementations that are possible by combining various elements to define more implementations. In some examples, some implementations may include fewer elements without departing from the spirit of the disclosed or equivalent implementations.
[0091] Embodiments of the present disclosure may be practiced with a variety of computer system configurations, including handheld devices, microprocessor systems, microprocessor-based or programmable consumer electronics, minicomputers, mainframe computers, etc. Embodiments of the present disclosure may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a wire-based or wireless network.
[0092] With the foregoing embodiments in mind, it will be appreciated that embodiments of the present disclosure may employ various computer-implemented operations involving data stored in computer systems. These operations are operations requiring physical manipulation of physical quantities. Any of the operations described herein that form part of the embodiments of the present disclosure are useful machine operations. Embodiments of the present disclosure also relate to devices or apparatus for performing these operations. The apparatus may be specially constructed for the required purposes, or the apparatus may be a general-purpose computer selectively activated or configured by a computer program stored in the computer. In particular, various general-purpose machines may be used with computer programs written in accordance with the teachings herein. Alternatively, it may prove more convenient to construct a more specialized apparatus to perform the required operations.
[0093] The present disclosure can also be embodied as computer readable code on a computer readable medium. The computer readable medium can be any data storage device that can store data. The data can then be read by a computer system. Examples of computer readable media include hard drives, network attached storage (NAS), read only memory, random access memory, CD-ROM, CD-R, CD-RW, magnetic tape, and other optical and non-optical data storage devices. The computer readable medium can include computer readable tangible media distributed across network connected computer systems such that the computer readable code is stored and executed in a distributed fashion.
[0094] Although the method operations have been described in a particular order, it should be understood that other housekeeping operations may be performed between the operations, or the operations may be adjusted to occur at slightly different times, or may be distributed within a system that allows processing operations to occur at various intervals related to processing, so long as the processing of the overlay operation is performed in the desired manner.
[0095] Although the foregoing disclosure has been described in some detail for clarity of understanding, it will be apparent that certain changes and modifications can be practiced within the scope of the appended claims. Thus, the present embodiments are to be considered as illustrative and not restrictive, and the embodiments of the present disclosure are not to be limited to the details provided herein, but may be modified within the scope and equivalents of the appended claims.
Claims
1. Tracking the state of a plurality of compute nodes in a data center, the plurality of compute nodes being distributed across a plurality of rack assemblies; Track multiple popularity ratings for multiple video games, determining a distribution of one or more video games from the plurality of video games across the plurality of rack assemblies based on the plurality of popularity ratings of the plurality of video games; storing the one or more video games from the plurality of video games across the plurality of rack assemblies based on the determined distribution of the one or more video games.
2. The tracking of the states of the plurality of computing nodes comprises: Tracking utilization of each of the plurality of compute nodes; The method of claim 1 , further comprising tracking distribution of the plurality of video games across a plurality of network storage of the plurality of rack assemblies.
3. Additionally, it uses artificial intelligence to determine trends in video game popularity ratings, The method of claim 1 , further comprising determining a number of rack assemblies for storing the video games based on the trend.
4. In said tracking of said plurality of popularity ratings, receiving telemetry data from a plurality of sites accessible via one or more networks; Extracting features from the telemetry data; classifying a trend of the video game based on the plurality of features, the trend being indicative of an increasing popularity rating of the video game, the popularity rating corresponding to demand for game sessions of the video game; The method of claim 1 , comprising increasing a number of rack assemblies of the plurality of rack assemblies for storing the video games based on the trend.
5. said tracking of said plurality of population assessments further comprising: The method of claim 1 , further comprising tracking one or more game sessions of each of the plurality of video games across the plurality of compute nodes of the data center.
6. further comprising: tracking one or more game sessions of a video game across the plurality of compute nodes of the data center; identifying a trend based on the one or more tracked game sessions of the video game that indicates an increasing popularity rating of the video game based on the one or more tracked game sessions of the video game, the popularity rating corresponding to demand for game sessions of the video game; The method of claim 5 , further comprising increasing a number of rack assemblies of the plurality of rack assemblies for storing the video games based on the trend.
7. tracking a game session of a video game across the plurality of compute nodes of the data center; predicting periods of high demand for game sessions of said video game; 6. The method of claim 5, further comprising loading the video games across a number of rack assemblies of the plurality of rack assemblies to support the high demand during the period of time.
8. tracking a first history of a first popularity rating of the first video game across the plurality of compute nodes of the data center for a portion of the existence of the first video game; determining a first trend in the first history indicating that a first popularity rating of the first video game has increased over a first period of time during the portion of the existence of the first video game; determining a first number of rack assemblies for storing the first video game based on the first trend; predicting a second trend indicating that a second popularity rating for a second video game over a second time period will be similar to the first trend for the first video game; 2. The method of claim 1, further comprising loading a second number of rack assemblies for storing said second video games based on said second trend, said second number being approximately the same as said first number.
9. further comprising: identifying a trend indicative of an increasing popularity rating of a video game based on said tracking of said popularity ratings of said plurality of video games; The method of claim 1 , further comprising: preloading the video game onto one or more computing nodes of the plurality of computing nodes for a new game session based on the trend.
10. A non-transitory computer readable medium storing a computer program for performing a method, the non-transitory computer readable medium comprising: program instructions for tracking a state of a plurality of computing nodes in a data center, the plurality of computing nodes being distributed across a plurality of rack assemblies; having program instructions for tracking a plurality of popularity ratings for a plurality of video games; determining a distribution of one or more video games from the plurality of video games across the plurality of rack assemblies based on the plurality of popularity ratings of the plurality of video games; A non-transitory computer-readable medium having program instructions for storing the one or more video games from the plurality of video games across the plurality of rack assemblies based on the determined distribution of the one or more video games.
11. The program instructions for tracking the state of the plurality of compute nodes include: program instructions for tracking utilization of each of the plurality of compute nodes; and program instructions for tracking distribution of the plurality of video games across a plurality of network storages of the plurality of rack assemblies.
12. program instructions for using artificial intelligence to determine trends in popularity ratings of video games; 11. The non-transitory computer readable medium of claim 10, further comprising: program instructions for determining a number of rack assemblies for storing the video games based on the trend.
13. The program instructions for tracking a plurality of popularity ratings include: having program instructions for receiving telemetry data from a plurality of sites accessible via one or more networks; having program instructions for extracting features from the telemetry data; and program instructions for classifying a trend of the video game based on the plurality of characteristics, the trend being indicative of an increasing popularity rating of the video game, the popularity rating corresponding to demand for game sessions of the video game.
11. The non-transitory computer readable medium of claim 10 having program instructions for increasing a number of rack assemblies of the plurality of rack assemblies for storing the video games based on the trend.
14. further comprising program instructions for tracking one or more game sessions of a video game across the plurality of compute nodes of the data center; and identifying a trend based on the one or more tracked game sessions of the video game that indicates an increasing popularity rating of the video game based on the one or more tracked game sessions of the video game, the popularity rating corresponding to demand for game sessions of the video game; 11. The non-transitory computer readable medium of claim 10 having program instructions for increasing a number of rack assemblies of the plurality of rack assemblies for storing the video games based on the trend.
15. program instructions for tracking a first history of a first popularity rating of a first video game across the plurality of compute nodes of the data center for a portion of the existence of the first video game; determining a first trend in the first history indicating that a first popularity rating of the first video game has increased over a first period of time during the portion of the existence of the first video game; determining a first number of rack assemblies for storing the first video game based on the first trend; and program instructions for predicting a second trend indicating that a second popularity rating for a second video game over a second time period will be similar to the first trend for the first video game; 11. The non-transitory computer readable medium of claim 10, having program instructions for loading a second number of rack assemblies for storing the second video games based on the second trend, the second number being approximately the same as the first number.
16. 1. A computer system comprising: A processor is included. a memory coupled to the processor and storing instructions which, when executed by the computer system, cause the computer system to perform a method, the method comprising: Tracking the state of a plurality of compute nodes in a data center, the plurality of compute nodes being distributed across a plurality of rack assemblies; Track multiple popularity ratings for multiple video games, determining a distribution of one or more video games from the plurality of video games across the plurality of rack assemblies based on the plurality of popularity ratings of the plurality of video games; and and storing the one or more video games from the plurality of video games across the plurality of rack assemblies based on the determined distribution of the one or more video games.
17. In the method, the tracking of the states of the plurality of computing nodes comprises: Tracking utilization of each of the plurality of compute nodes; 20. The computer system of claim 16, further comprising: a computer system configured to track distribution of the plurality of video games across a plurality of network storage devices of the plurality of rack assemblies.
18. In the method, the tracking of the plurality of popularity ratings comprises: receiving telemetry data from a plurality of sites accessible via one or more networks; Extracting features from the telemetry data; classifying a trend of the video game based on the plurality of features, the trend being indicative of an increasing popularity rating of the video game, the popularity rating corresponding to demand for game sessions of the video game; 17. The computer system of claim 16, further comprising: increasing a number of rack assemblies of the plurality of rack assemblies for storing the video games based on the trend.
19. In the method, tracking one or more game sessions of a video game across the plurality of compute nodes of the data center; identifying a trend based on the one or more tracked game sessions of the video game that indicates an increasing popularity rating of the video game based on the one or more tracked game sessions of the video game, the popularity rating corresponding to demand for game sessions of the video game; 17. The computer system of claim 16, further comprising: increasing a number of rack assemblies of the plurality of rack assemblies for storing the video games based on the trend.
20. In the method, identifying a trend indicative of an increasing popularity rating of a video game based on said tracking of the popularity ratings of the plurality of video games; 2. The computer system of claim 1, further comprising preloading the video game onto one or more computing nodes of the plurality of computing nodes for a new game session based on the trend.
Citation Information
Patent Citations
Predictive instant play for an application over the cloud
EP3791943A1
Information processing system, information processing apparatus, scaling method, program, and recording medium
JP2012208781A
Server management device and server management method
JP2017037445A
Information processing system, and information processing program
JP2017113375A
Machine learning driven resource allocation
US20200171382A1
