Cloud broadcast system and method based on multi-channel and multi-dimensional resource elastic allocation

Abstract

The present invention relates to a cloud broadcasting system and method based on elastic allocation of multi-channel and multi-dimensional resources, including: a time dimension condition collector, a security dimension condition collector, and a business dimension condition collector, which are connected in sequence to a dimension condition analyzer and a physical server cluster. An elastic expansion manager and a physical network card with SR-IOV are provided in the physical server cluster, and a virtual machine group consisting of multiple virtual machines is virtualized in the physical server according to the requirements of the elastic expansion manager. The present invention introduces the time dimension, security dimension, and business dimension into the elastic scaling conditions of the broadcasting service, and uses a dimension condition analyzer to analyze and process the multi-dimensional scaling conditions, and clarifies the conditions that need to trigger the elastic scaling of the broadcasting service, thereby improving the flexibility and security of the broadcasting service. At the same time, the single-root I / O virtualization technology is used for the IP signal broadcasting server, and the cloud virtual machine drift function is fully utilized to realize the flexibility expansion of the broadcasting service so that each virtual machine IP signal is independent.

Description

Technical Field

[0001] The present invention relates to a cloud broadcasting system and method based on multi-channel and multi-dimensional resource elastic allocation, an electrical processing system and method of a digital network, and a device and method for broadcast and television broadcast control. Background Art

[0002] In the TV station broadcasting system, the broadcast server is responsible for broadcasting the program material files or external real-time signals to be broadcast in sequence according to the time arrangement plan of the program list. At present, the broadcasting system generally has the problem of high system construction cost and insufficient business flexibility. On the one hand, the broadcast server is generally an expensive rack-mounted physical server with high configuration. In order to ensure the security and stability of the radio and television broadcasting business, each broadcast channel of the broadcasting system will be equipped with more than two broadcast servers working simultaneously as hot standby for each other, which leads to high cost of the broadcast servers in the broadcasting system. On the other hand, because the rack-mounted physical servers usually used by the broadcast server do not have elastic expansion capabilities, when the business scale of the broadcast system changes, the performance boundary of the broadcast server cannot be expanded on demand, which reduces business flexibility and may also lead to waste of system resources during idle time, further reducing the cost-effectiveness of system construction.

[0003] At present, most broadcast systems are built with two signal links, one main and one backup. Each link contains a broadcast server. The two links, one main and one backup, work simultaneously and jointly undertake the broadcast tasks of three to four channels. The construction and use mode of this system determines that a single system will serve the broadcast tasks of three to four channels for a long time. When business demand expands, the system cannot quickly expand its processing capacity to cope with the growing business demand, resulting in insufficient business flexibility of the system; when business demand shrinks, system resources cannot be released, resulting in waste of resources in view of the high cost of broadcast servers. On the other hand, after the broadcast business is moved to the cloud, the trigger conditions for the elastic scaling of the system boundary of the current broadcast method are mostly failures in the business system, and it cannot respond to the increase or decrease in business demand. Although the redundancy of the system is improved, the flexibility of the system is not effectively improved. Summary of the invention

[0004] In order to overcome the problems of the prior art, the present invention proposes a cloud broadcast system and method based on multi-channel and multi-dimensional resource elastic allocation. The system and method can arrange the use of virtual machine resources reasonably and effectively according to the needs of broadcast business dimensions, time dimensions, etc., provide clear resource scaling guidance basis and system scaling trigger conditions that meet business characteristics for broadcast business on the cloud, and further improve the flexibility and security of broadcast business.

[0005] The object of the present invention is achieved as follows: a cloud broadcast system based on multi-channel multi-dimensional resource elastic allocation, comprising: a time dimension condition collector, a security dimension condition collector, and a business dimension condition collector; the time dimension condition collector, the security dimension condition collector, and the business dimension condition collector are connected to a dimension condition analyzer and a physical server cluster in sequence; the physical server cluster is provided with an elastic expansion manager and a physical network card with SR-IOV, and a virtual machine group consisting of multiple virtual machines is virtualized in the physical server according to the requirements of the elastic expansion manager, each of the virtual machines is provided with its own virtual IO driver, and the physical network card is provided with virtual IO corresponding to the number of virtual machines; the virtual machine group is connected to the time dimension condition collector, the security dimension condition collector, and the business dimension condition collector;

[0006] The time dimension condition collector is used to collect changes in demand for resource quantities due to changes in the state of the broadcast service over time;

[0007] The security dimension condition collector is used to monitor the resource usage and working status of each virtual machine to determine whether the relevant services of each virtual machine are in normal operation;

[0008] The business dimension condition collector is used to collect changes in the demand for resource quantity of broadcasting due to changes in business status;

[0009] The dimension condition analyzer is used to determine whether it is necessary to increase or decrease virtual servers and corresponding virtual IOs according to the dimension conditions;

[0010] The elastic expansion manager is used to start or shut down the virtual machine according to the broadcast status analyzed by the dimension condition analyzer;

[0011] The virtual machine is used to set at least one broadcast channel for broadcast services for use by one or more broadcast channels.

[0012] A cloud broadcast method using the above system for flexible resource allocation, the steps of the method are as follows:

[0013] Step 1, initiate broadcast: According to the broadcast plan, the dimension condition analyzer calculates the number of virtual machines to be used, the number of elastic expansion manager virtual machines starts virtualization and enables each virtual machine to build a corresponding number of input and output channels with the SR-IOV physical network card, and then starts normal broadcasting business;

[0014] Step 2, monitoring and collecting information: During normal broadcasting, the time dimension condition collector, security dimension condition collector, and business dimension condition collector monitor the broadcasting status, and respectively collect business dimension conditions, security dimension conditions, and time dimension conditions during the broadcasting process, and send them to the dimension condition analyzer;

[0015] Step 3, analyzing the working status of the virtual machine: the dimension condition analyzer analyzes the working status of each virtual machine according to the received business dimension conditions, security dimension conditions, and time dimension conditions, and determines whether the virtual machine needs to be turned on or off. If yes, proceed to the next step, otherwise return to step 2 to continue monitoring the broadcast process;

[0016] Step 4, start or shut down the virtual machine: the elastic expansion manager starts or shuts down the virtual machine according to the calculation analysis of the dimension condition analyzer;

[0017] Step 5: Modify the input and output channels: The elastic expansion manager increases or decreases the input and output channels according to the number of virtual machines that are turned on or off;

[0018] Step 6, service drift: drift the broadcast service running on other virtual machines to the newly started virtual machine, or drift the broadcast service on the shut down virtual machine to other virtual machines, and return to step 2 to continue monitoring the broadcast process.

[0019] The advantages and beneficial effects of the present invention are as follows: the present invention introduces the time dimension, security dimension, and business dimension into the elastic scaling conditions of the broadcasting business, and uses a dimensional condition analyzer to analyze and process the multi-dimensional scaling conditions, thereby clarifying the conditions that need to trigger the elastic scaling of the broadcasting business, thereby improving the flexibility and security of the broadcasting business, and realizing the on-demand use of resources, and reducing the cost of system use. At the same time, the single-root I / O virtualization technology is used for the IP signal broadcasting server, which can not only fully utilize the cloud virtual machine drift function to realize the flexible expansion of the broadcasting business, but also ensure the independence of the IP signal traffic on each virtual machine, ensure the stability of the broadcasting system, make it possible for the broadcasting business to go to the cloud to enhance the flexibility of the broadcasting business, and fully utilize the elastic expansion capability of the cloud platform to reduce the construction cost when the system business volume boundary changes and the maintenance cost when the system fails. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0021] Figure 1 is a principle block diagram of the system described in Embodiment 1 of the present invention;

[0022] Figure 2 It is a flow chart of the method described in Embodiment 1 of the present invention. DETAILED DESCRIPTION

[0023] Embodiment 1:

[0024] This embodiment is a cloud broadcast system based on multi-channel and multi-dimensional resource elastic allocation. Figure 1 As shown. This embodiment includes: a time dimension condition collector, a security dimension condition collector, and a business dimension condition collector; the time dimension condition collector, the security dimension condition collector, and the business dimension condition collector are connected to the dimension condition analyzer and the physical server cluster in sequence; the physical server cluster is provided with an elastic expansion manager and a physical network card with SR-IOV, and the physical server is virtualized into a virtual machine group consisting of multiple virtual machines according to the requirements of the elastic expansion manager, each of the virtual machines is provided with its own virtual IO driver, and the physical network card is provided with virtual IO corresponding to the number of virtual machines; the virtual machine group is connected to the time dimension condition collector, the security dimension condition collector, and the business dimension condition collector.

[0025] In order to put the broadcast server on the cloud, there are two paths to achieve the existing technology. One is to establish a virtual machine on the physical server as a broadcast server one-to-one, and the physical machine network card and the virtual machine network card are connected one-to-one to transmit IP broadcast signals. Since the broadcast service requires continuity and is a lag-sensitive service, the problem with this method is that since the physical machine and the virtual machine are in a strict one-to-one relationship, the drift function in the blade box of the cloud server cannot be used. When the virtual machine fails, it is necessary to open another physical machine and establish a virtual machine, and then wait for the system initialization to complete before migrating the service. The timeliness is low, which affects the continuity of the broadcast service; the other is to establish multiple virtual machines on the physical machine one-to-many, and each virtual machine shares a common network card of the physical machine for IP signal transmission. When a single virtual machine fails, this method can seamlessly transfer the service to another virtual machine on the physical machine where the virtual machine is located through the drift function in the physical machine, ensuring the continuity of the broadcast service. However, since multiple virtual machines share a network card, when the business volume fluctuates, it may cause the network port traffic of each virtual machine to occupy each other, thereby causing network traffic fluctuations, which in turn affects the jitter of the broadcast IP signal and causes the instability of the broadcast service.

[0026] This embodiment uses a one-to-many method of establishing multiple virtual machine groups (cloud servers) on a physical server cluster (physical machine), so that the virtual machine group used as a broadcast server can achieve drift between virtual machines in an emergency. At the same time, in order to avoid the business instability caused by multiple virtual machines directly sharing a network card in this method, the single-root I / O virtualization technology SR-IOV is used to divide the function and performance of the physical network card, and the traffic bypasses the virtual switch, so that each virtual machine has exclusive performance of a specific scale of the physical network card, eliminating the possibility of data jitter caused by multiple virtual machines sharing a single network card.

[0027] After the broadcast service is on the cloud, the trigger conditions for elastic scaling of the system boundary in the existing technical methods are mostly failures in the business system, and they cannot respond to the increase or decrease of business demand. Although the redundancy of the system is improved, the flexibility of the system is not effectively improved. This embodiment uses the time dimension, security dimension, and business dimension as trigger conditions for elastic scaling at the same time. After adding a dimensional condition analyzer to make a comprehensive judgment on the multi-dimensional conditions, the elastic expansion manager of the cloud server issues a final elastic expansion or contraction instruction, so that the virtual machine group can not only achieve redundant security through automatic elastic expansion in the event of a failure, but also automatically elastically scale according to the real-time busyness of the system and the business plan of the system in different time periods. By using it on demand, the system usage cost is reduced while the system efficiency is improved.

[0028] The security dimension condition collector is used to monitor the resource usage and working status of each virtual machine to determine whether the services related to each virtual machine are in normal operation. The security dimension condition refers to determining whether the services related to the virtual machine are in normal operation by monitoring the resource usage and working status of each virtual machine in the system, such as whether the network traffic is stable, whether the CPU / memory usage is too high, whether the broadcast signal meets the technical standards, etc. When any of the above indicators is abnormal, it is considered that the security dimension condition of the broadcast service on the virtual machine has changed, and it is necessary to send trigger condition data to the dimension condition analyzer.

[0029] The business dimension condition collector is used to collect the changes in the demand for resources due to changes in business status. Business dimension conditions refer to changes in the business status of the system, which causes changes in the system's demand for resources. For example, the original broadcast channel belongs to an ordinary channel, which only contains two main and backup broadcast IP signal links. Due to changes in business needs, the channel is changed to a key broadcast channel (such as for major occasions such as the Spring Festival Gala live broadcast), which needs to include three main, backup, and triple backup broadcast IP signal links. At this time, more broadcast server instances are needed to open more broadcast channels, that is, it is considered that the business dimension conditions of the broadcast business on the virtual machine have changed, and it is necessary to send trigger condition data to the dimension condition analyzer.

[0030] The time dimension condition collector is used to collect the changes in the demand for resource quantity due to the change of the state of the broadcast business over time. The time dimension condition means that according to the arrangement of the time schedule, the business state of the virtual machine group responsible for broadcasting will change the demand for resource quantity over time. For example, according to the broadcast schedule, a broadcast channel broadcasts live signals during the day, requiring three external broadcast channels, and replays material files at night, without the need for external broadcast channels. At this time, it is necessary to expand or shrink the broadcast server instance according to the time schedule to ensure that the business needs are met while minimizing resource waste and reducing the cost of system use. That is, it is considered that the time dimension condition of the broadcast business on the virtual machine has changed, and it is necessary to send trigger condition data to the dimension condition analyzer.

[0031] At the same time, this embodiment also applies single-root I / O virtualization technology to the network design of a single physical machine to multiple virtual machines by setting up a physical network card with SR-IOV to ensure the stability of the business. By deploying multi-channel broadcast servers on the cloud, the redundant security of the system is increased, and the elastic expansion of the system is triggered by multi-dimensional business conditions, the flexibility of the business is achieved, and the changes in the number of virtual machines caused by changes in the broadcast business volume are responded to in real time, achieving the best resource utilization.

[0032] Single Root I / O Virtualization technology SR-IOV is a specification that enables a single Peripheral Component Interconnect Express (PCIe) physical device under a single root port to appear as multiple separate physical devices to the hypervisor or guest operating system. Virtual machines can use SR-IOV virtual functions for network connections. The virtual machine and the physical adapter exchange data directly without using a virtual switch as an intermediary. Bypassing the virtual switch for network connections reduces latency and improves CPU efficiency.

[0033] After the virtual machine uses single-root I / O virtualization technology to ensure the exclusivity of multi-network card performance, multiple broadcast channels are established on the virtual machine as the primary and backup links of the broadcast channels. Each broadcast channel is divided into 2 IP signal input channels and 1 IP signal output channel. Through business volume calculation, the IP traffic of each channel is divided and bound with the exclusive traffic obtained by the virtual machine through SR-IOV, so as to achieve stability and redundancy drift of virtual broadcast services after going to the cloud.

[0034] The dimension condition analyzer is used to determine whether it is necessary to increase or decrease virtual servers and corresponding virtual IOs according to the conditions of each dimension. After receiving the multi-dimensional condition data, the dimension condition analyzer first locks the most basic resource requirements according to the current security dimension conditions, and then makes resource allocation and time scheduling plans for resource allocation according to business conditions and time conditions.

[0035] The elastic expansion manager is used to start or shut down the virtual machine according to the broadcast status analyzed by the dimension condition analyzer. After analysis and calculation by the dimension condition analyzer, the dimension condition analyzer outputs instructions to the elastic expansion manager in the physical server cluster, and the elastic expansion manager starts or shuts down the virtual machine according to the instructions, thereby ensuring that business needs are met while minimizing resource waste and reducing system usage costs.

[0036] The virtual machine is used to set up at least one broadcast channel for broadcast services for use by one or more broadcast channels. Figure 1 As shown, there are multiple virtual machines in the virtual machine group, and each virtual machine has multiple broadcast channels, such as Figure 1 As shown, broadcast channel 1 of virtual machine 1 and broadcast channel 1 of virtual machine 2 jointly undertake the broadcast of broadcast channel A, and broadcast channel 2 of virtual machine 1 and broadcast channel 2 of virtual machine 2 jointly undertake the broadcast of broadcast channel B.

[0037] Embodiment 2:

[0038] The present embodiment is a cloud-based broadcasting method for elastic resource allocation of the system described in the embodiment. According to the requirements of the broadcasting business dimension, time dimension, and security dimension, the present embodiment arranges the use of virtual machine resources in a reasonable and effective manner in a staggered manner, provides a clear resource scaling guidance basis and system scaling trigger conditions that meet the business characteristics for the broadcasting business on the cloud, and improves the flexibility and security of the broadcasting business. The specific approach is to monitor the needs of the broadcasting business, and adjust the number of virtual machines in real time according to the business, time, and security requirements to achieve the best configuration of the number of virtual machines and ensure full utilization of resources. The present embodiment is mainly for the broadcast processing of IP video signals.

[0039] The steps of the method described in this embodiment are as follows:

[0040] Step 1, initiate broadcast: According to the broadcast plan, the dimension condition analyzer calculates the number of virtual machines to be used, the elastic expansion manager starts the virtualization and enables each virtual machine to build a corresponding number of input and output channels with the SR-IOV physical network card, and then starts normal broadcasting business.

[0041] When a broadcast plan appears, the first thing to do is to calculate how many broadcast resources are needed for this broadcast plan to achieve normal broadcast and the corresponding redundant virtual machines (virtual broadcast servers). Therefore, it is necessary to calculate and analyze according to the broadcast plan, find the optimal number of virtual machines and the corresponding number of input and output channels, establish the corresponding virtual IO, and then start normal broadcast.

[0042] Step 2, monitor and collect information: During the normal broadcast process, the time dimension condition collector, security dimension condition collector, and business dimension condition collector monitor the broadcast status, and respectively collect the business dimension conditions, security dimension conditions, and time dimension conditions during the broadcast process, and send them to the dimension condition analyzer.

[0043] This step is to monitor whether the virtual server group during the broadcast process is suitable for the current broadcast business and whether any broadcast problems occur.

[0044] Step 3, analyze the working status of the virtual machine: The dimension condition analyzer analyzes the working status of each virtual machine based on the received business dimension conditions, security dimension conditions, and time dimension conditions, and determines whether the virtual machine needs to be turned on or off. If so, proceed to the next step, otherwise return to step 2 to continue monitoring the broadcast process.

[0045] This step is to determine any problems that may occur during the broadcast process, such as: whether the demand for resource volume changes over time, whether the broadcast business increases, whether the broadcast is hindered due to resource congestion, or conversely, some virtual machines are not fully utilized, resulting in resource shortage, etc., and determine whether it is necessary to increase or decrease the resources of the virtual server based on the cause of the problem.

[0046] Step 4, start or shut down the virtual machine: The elastic expansion manager starts or shuts down the virtual machine based on the calculation analysis of the dimension condition analyzer.

[0047] The activation of the virtual machine is to add a virtual server in the physical server cluster, while the deactivation of the virtual machine is the opposite, that is, to deactivate the virtual server that has been virtualized to save hardware resources.

[0048] Step 5: Modify input and output channels: The elastic expansion manager increases or decreases input and output channels according to the number of virtual machines that are turned on or off.

[0049] Each virtual machine requires its own IO channel. Therefore, when building a virtual machine, it is necessary to install a virtual IO driver on the virtual machine and create a corresponding dedicated IO channel on the physical network card with SR-IOV to improve the output efficiency and flexibility of the virtual machine.

[0050] Step 6, service drift: drift the broadcast service running on other virtual machines to the newly started virtual machine, or drift the broadcast service on the shut down virtual machine to other virtual machines, and return to step 2 to continue monitoring the broadcast process.

[0051] If a new virtual machine is created, a portion of the broadcast business of the original virtual machine will be allocated to the newly built virtual machine. If the virtual machine is shut down, the broadcast business on the shut down virtual machine will be moved to other virtual machines that are working but have less business, so as to maximize broadcast security and broadcast resource efficiency.

[0052] Finally, it should be noted that the above is only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred arrangement scheme, a person skilled in the art should understand that the technical solution of the present invention (such as the form of the broadcast system, the form of the physical server cluster and the virtual machine group, the sequence of steps, etc.) can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A cloud broadcast system based on multi-channel and multi-dimensional resource elastic allocation, characterized by: include: A time dimension condition collector, a security dimension condition collector, and a business dimension condition collector; the time dimension condition collector, the security dimension condition collector, and the business dimension condition collector are connected to the dimension condition analyzer and the physical server cluster in sequence; the physical server cluster is provided with an elastic expansion manager and a physical network card with SR-IOV, and the physical server is virtualized into a virtual machine group consisting of multiple virtual machines according to the requirements of the elastic expansion manager, each of the virtual machines is provided with its own virtual IO driver, and the physical network card is provided with virtual IO corresponding to the number of virtual machines; the virtual machine group is connected to the time dimension condition collector, the security dimension condition collector, and the business dimension condition collector; The time dimension condition collector is used to collect changes in resource demand due to changes in the state of the broadcast service over time; The security dimension condition collector is used to monitor the resource usage and working status of each virtual machine to determine whether the relevant services of each virtual machine are in normal operation; The business dimension condition collector is used to collect changes in broadcast resource requirements due to changes in business status; The dimension condition analyzer is used to determine whether it is necessary to increase or decrease virtual servers and corresponding virtual IOs according to the dimension conditions; The elastic expansion manager is used to virtualize or shut down a virtual machine according to the broadcast status analyzed by the dimension condition analyzer; The virtual machine is used to set at least one broadcast channel for broadcast services for use by one or more broadcast channels.

2. A cloud broadcast method using the cloud broadcast system based on multi-channel and multi-dimensional resource elastic allocation according to claim 1, characterized in that: The steps of the method are as follows: Step 1: Initiate broadcast: Based on the broadcast plan, the dimension condition analyzer calculates the number of virtual machines to be used. The elastic expansion manager starts virtual machines based on the number of virtual machines and establishes a corresponding number of input and output channels between each virtual machine and the SR-IOV physical network card. Then, normal broadcast services begin. Step 2, monitoring and collecting information: During normal broadcasting, the time dimension condition collector, security dimension condition collector, and business dimension condition collector monitor the broadcasting status, and respectively collect business dimension conditions, security dimension conditions, and time dimension conditions during the broadcasting process, and send them to the dimension condition analyzer; Step 3: Analyze the working status of the virtual machine: The dimension condition analyzer analyzes the working status of each virtual machine based on the received business dimension conditions, security dimension conditions, and time dimension conditions, and determines whether the virtual machine needs to be turned on or off. If yes, it proceeds to the next step; otherwise, it returns to step 2 to continue monitoring the broadcast process; Step 4: Start or shut down the virtual machine: The elastic expansion manager starts or shuts down the virtual machine based on the calculation and analysis of the dimension condition analyzer; Step 5: Modify input and output channels: The elastic expansion manager increases or decreases input and output channels according to the number of virtual machines that are turned on or off; Step 6, service drift: drift the broadcast services currently running on other virtual machines to the newly started virtual machine, or drift the broadcast services on the shut down virtual machine to other virtual machines, and return to step 2 to continue monitoring the broadcast process.

Images