Sticky Routing via Unique ID Extraction in Server Farms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for handling client requests in server farms face challenges in efficiently routing requests while maintaining session information and balancing server loads, which can expose internal server farm configurations, posing security risks and administrative burdens.
Innovation Solution
Upstream devices, such as load balancers, dynamically generate unique IDs to route requests within the server farm, enabling sticky routing by embedding or encoding these IDs within requests, allowing for secure and efficient session management without revealing internal configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If session IDs are used to route requests to specific servers, then session information can be maintained on a single server, but the mapping data size becomes prohibitive and security risks increase
Solution Approach 1:
The patent extracts the server identification information from the session ID itself. Instead of using the full session ID for routing, a separate unique ID (extracted or derived from the session ID) is used to identify the target server. This separates the session management function from the routing function, reducing the mapping data size while maintaining reliable session information maintenance.
Solution Approach 2:
The patent introduces a unique ID as an intermediary between the session ID and the server identification. This unique ID serves as a compact key that can be used by load balancers to route requests without requiring large mapping tables. The unique ID acts as a mediator that links the session information to the specific server while minimizing data storage requirements.
2Ease of operation
If server addresses are exposed to clients for routing, then standard Internet routing protocols can be used, but internal server farm configuration information is revealed creating security risks
Solution Approach 1:
The patent creates a simplified copy of the server identification system. Instead of exposing real server addresses and internal configuration details, the system uses abstract unique IDs that represent servers without revealing their actual identities or the farm's internal structure. These unique ID copies enable routing operations while masking the true configuration from external clients.
Solution Approach 2:
The unique ID serves as an intermediary that shields the internal server farm configuration from external exposure. Clients interact with abstract unique IDs rather than real server addresses, and the load balancer translates these IDs into internal routing information. This intermediary layer maintains ease of routing operation while preventing security risks associated with configuration exposure.
3Reliability
If load balancers maintain detailed session mappings, then sticky routing can be implemented, but administrative burden increases
Solution Approach 1:
The patent extracts only the essential routing information (unique ID) from the full session mapping data. Load balancers maintain minimal mapping tables that associate unique IDs with server identifiers, rather than storing complete session information. This extraction approach enables sticky routing functionality while significantly reducing the administrative burden of managing large mapping datasets.
Solution Approach 2:
The system enables load balancers to automatically generate and manage unique IDs without requiring manual configuration or intervention. The unique ID generation and mapping processes are automated, allowing the system to maintain sticky routing capabilities while minimizing administrative overhead. The load balancers self-manage the mapping data with reduced human intervention.
Data Source
AI summary
An upstream device, such as a load balancer or router within a server farm, may perform sticky routing of session requests to the servers handling those sessions by generating unique ID(s) identifying how requests may be routed through the server farm. Upstream devices through which travels a request that is not associated with a session on a server of the server farm may generate one or more unique IDs identifying how the request is routed through the server farm. The server handling the request may form a session ID and return that session ID and the unique ID(s) to the client that originated the new request and session. Clients may then send the session ID and unique IDs with clients requests for that session. Upon receiving requests corresponding to established sessions, the upstream devices may then identify routing information from the unique ID(s) and route the request accordingly.