Multiple Server Clustering Solutions
Our network topology consultants have years of experience designing and deploying complex architectures for small business, government, ISV's and SaaS companies. We work closely with you to ensure that your hosted infrastructure meets and exceeds your business requirements for traffic fluctuations, application performance, storage, disaster recovery, and security.
Here are some are some typical deployments we have experience with:
Server Cluster Scenarios
Server Cluster Scenarios
Scenario 1
High traffic website with read intensive database requests
- Redundant Load balancing appliance (including session persistence)
- Redundant firewalls (shared)
- 2x Web servers
- 1 database server (MSSQL, MySQL, Postgres)
- Back up (daily incremental, weekly full, 28 day retention)
Back to top
Scenario 2
High traffic website with redundant web and application servers, and database
- Redundant load balancer for application and Web servers
- Redundant dedicated firewalls
- 2 x Web servers
- 2 x application servers
- 2 x database servers (active/passive failover configuration)
- Server monitoring
- Backup (daily incremental, weekly full, 28 day retention) for each server
Back to top
Scenario 3
Mission critical business application with high performance, high availability, and disaster recovery requirements (SaaS/ASP providers ISV's)
- Redundant load balancers for web/application servers
- Redundant dedicated firewalls
- Web cluster 3x web servers
- Application cluster 3x by application servers
- Fiber channel storage area network
- Internal switch
- Application delivery and acceleration appliance
- Intrusion detection and prevention
- Off site tape back up
Back to top
The diagram below describes a fairly standard server cluster. In this example, the users are accessing a single Master Control Server that directs the request to the servers within the cluster. The dedicated servers in the cluster are acting as a single unit, sharing the application as well as the runtime state for that application. This allows for a seamless failover if an individual server fails, because the state is maintained within the cluster.
Server clustering is different from server load balancing. In the above example, although state is maintained in the application and allows for redundancy within a session, it still requires a single server, the Master Control Server, to be online. If this server fails, the whole cluster fails.
