In reading a very useful book for those wanting to know mare about the force.com platform, I’ve taken this excerpt that really gives good information about the infrastructure that Advologix applications run on. For further reading I HIGHLY recommend :
A newcomer’s guide to building applications on Salesforce.com and the Force.com platform
by Jeff Douglas & Wes Nolte
Infrastructure is one of those things that you sometimes take for granted in a “Cloud” environment. Development in the cloud inherently removes the necessity to be concerned about infrastructure and allows us to focus more resources on business processes. However, infrastructure is one of the secret ingredients of salesforce.com. Over 77,000 companies and commercial vendors trust the platform to deliver robust, reliable, Internet-scale applications.
Salesforce.com has five main data centers and is currently building out two new ones in the US. Equinix owns these facilities and salesforce.com co-locates their servers there. To ensure availability and redundancy, salesforce.com uses multiple network vendors at these locations.
| Location | Function |
|---|---|
| San Jose, CA | Main Production Data Center |
| Reston, VA | Production & Back-up Data Center |
| Singapore | Production Data Center |
| Japan | Production Data Center |
| San Francisco, CA | R&D Lab and Tape Archive |
| London, UK | Coming 2012 |
Each user that logs into salesforce.com is mapped to what is called an “Org”, a complete “instance” of the salesforce.com software stack. Salesforce.com has roughly 3000 Dell servers (50% are idle, staged for disaster recovery) configured in a “pod” infrastructure. These pods are clustered pools of RedHat Linux boxes running open source Resin app servers, Oracle RAC database instances, Java app servers, Lucene full-text search servers, load balancing servers and a SAN disk array. Requests to these app servers are stateless providing for a high degree of scalability and allowing for long running transactions to utilize multiple boxes. The salesforce.com network architecture is designed for massive failover. Virtually every transaction from the primary data centers are replicated in real-time to the Reston, VA back-up data center.
