Adaptive Data Center Routing for Cloud Gaming Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High performance network streaming applications like cloud gaming and cloud VR face performance issues due to inadequate network infrastructure, with traditional data center deployments and content delivery networks lacking sufficient resources and not accounting for specific application requirements, leading to latency, jitter, and packet loss problems.
Innovation Solution
Implementing a system that performs application-specific network tests to determine suitable data centers based on user device and network characteristics, optimizing data center distribution and forwarding to match application profiles, and using congestion control and resource utilization to ensure optimal performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If data centers are located far from user devices to provide wide coverage, then service availability is improved, but network latency increases making applications unplayable
Solution Approach 1:
The system segments the data center network into multiple regional data centers distributed across different geographic locations. Each data center serves a specific region, allowing users to be routed to the nearest appropriate data center. This segmentation resolves the contradiction by providing wide coverage through multiple locations while maintaining low latency within each regional segment.
Solution Approach 2:
The system adds a geographic dimension to data center deployment, transitioning from a single centralized data center to multiple distributed data centers across different regions. This dimensional change allows the system to simultaneously achieve wide service availability and low latency by matching users with geographically appropriate data centers.
2Ease of operation
If IP geolocation is used to route users to assigned data centers, then routing simplicity is improved, but application performance deteriorates due to distance and network infrastructure limitations
Solution Approach 1:
The system dynamically selects data centers based on real-time network performance measurements rather than static IP geolocation assignments. Network tests are performed to measure latency, jitter, and packet loss to specific data centers, and the system adapts routing decisions based on these dynamic conditions. This resolves the contradiction by maintaining routing simplicity while significantly improving application performance through performance-based selection.
Solution Approach 2:
The system implements feedback mechanisms where network performance metrics (latency, jitter, packet loss) are continuously measured and used to inform routing decisions. This feedback loop allows the system to identify and route users to data centers that provide optimal performance for specific applications, resolving the contradiction between simple routing and performance reliability.
3Productivity
If content delivery networks are used to support applications, then content delivery capability is improved, but computational resources for high-performance applications are insufficient
Solution Approach 1:
The system creates data centers that serve multiple functions: they provide content delivery capabilities like CDNs while simultaneously offering high-performance computational resources for demanding applications. By making data centers universal infrastructure that can handle both content delivery and compute-intensive workloads, the system resolves the contradiction between content delivery capability and computational resource availability.
4Speed
If data centers are assigned based on IP geolocation, then assignment speed is improved, but resource suitability for specific applications deteriorates
Solution Approach 1:
The system performs preliminary network tests and measurements to pre-establish performance characteristics between user devices and various data centers before application execution. This preliminary action creates a foundation of known performance data that enables rapid, informed routing decisions. The contradiction is resolved by preparing performance information in advance, allowing both fast assignment and application-specific resource suitability matching.
Data Source
AI summary
High performance applications—such as cloud game streaming, cloud virtual reality (VR), remote desktop, and others—are sensitive to various network conditions, such as latency, jitter, and packet loss. Systems of the present disclosure may match network characteristics for a user device with latency requirements for a particular application type, and application sessions may be forwarded or distributed to a suitable data center. To accomplish this, application specific network tests may be executed to determine requirements for executing a high performing application session for a user. The result of these tests, in addition to application specific performance requirements, may be used to find a suitable data center—from a set of available data centers—that is capable of hosting the application session without degradation. As a result, efficient use of distributed infrastructures may be accomplished, while avoiding congestion and hot-spots, and providing an optimized application experience for end users.