|Important||This document may not represent best practices for current development, links to downloads and other resources may no longer be valid. Current recommended version can be found here. ArchiveDisclaimer|
A Complex Team Foundation Server Topology
This topic describes a complex Team Foundation Server topology, with Team Foundation Server components installed to separate server nodes and accessed by clients in multiple domains in geographically distant physical locations. Team Foundation Server Proxy is installed in addition to additional fail-over components.
Complex Server Topology
The following diagram demonstrates a complex Team Foundation Server topology:
This example topology is similar to the moderate topology (see) in that Team Foundation Server services are deployed on one application-tier server and the Team Foundation Server databases are installed on a separate data-tier server, with Team Foundation Build and the team's test rig (agents) deployed on additional servers. Additionally, fail-over components have been added. An application-tier standby server has been added, as has a data tier with SQL clustering technologies.
In the example diagram, a child domain named Europe.Contoso.com represents a geographically distant child domain that uses a limited bandwidth connection. Since getting an enlistment from the source code control component of Team Foundation Server can take a very long time under such limited bandwidth conditions, a Team Foundation Server Proxy is installed in the child domain to act as a source file proxy cache. All client requests go directly to the application tier, with the exception of source code 'get' requests which are directed through the Team Foundation Server Proxy, which then caches all source files that it downloads.
The topology in this example is also concerned about fault tolerance, and has configured fail-over components. An application-tier standby server provides manual fail-over capabilities should the main application-tier server fail for any reason. Data tier fail-over is provided by taking advantage of the SQL Server 2005 fail-over features by using clustering.