GridGain is a powerful solution for web session clustering. HTTP is a stateless protocol so web servers are unaware of any relationship between consecutive requests. However, most web applications require some notion of a user session. The old approach of creating ‘sticky sessions’ by assigning users to specific servers in a cluster is usually not acceptable anymore, as it often causes uneven load distribution across a cluster with some servers overloaded while others are almost idle.
Cluster-aware and replicated web sessions enable load balancing of user sessions to any app server without losing intermediate user session data. With in-memory distributed caching technology, user sessions are available in memory regardless of which server in a cluster is affected because session state is replicated across multiple distributed servers. This approach:
- Provides automatic fault tolerance when a particular server crashes
- Supports out-of-the-box load balancing
- Offers high availability by avoiding a single point of failure since the session will still be available on another server
The GridGain In-Memory Data Fabric replicates session state across a cluster of distributed application servers, which enables the system to dynamically manage node failures. The GridGain In-Memory Data Fabric can be embedded in the application tier of existing applications without any modifications. It offers a high-performance web session clustering solution with advanced load balancing and no single point of failure. GridGain also offers unlimited scale with unparalleled economics on commodity hardware.