Welcome to MSDN Blogs Sign in | Join | Help

Microsoft's Team Foundation Server Deployment Topology

Given how much I talk about our internal "dogfooding" of TFS on this blog, I thought you might be interested in seeing a diagram of what our internal deployment actually looks like.  This diagram was put together by our IT department a couple of weeks ago.  It shows the basic network topology, regions we have servers and the number and purpose of various servers.  I don't even know what all of them are but I'll point out some highlights.

  • We have servers both on the intranet and the extranet.
  • There are servers and/or proxies in each of 5 different regions around the world and accessed from more places than that.  We're in the process of setting up a proxy in China right now but it wasn't in place at the time this diagram was made.
  • DevDiv - The server used by Developer Division (this is the server I publish all of my dogfood stats about).
  • MSIT - The server used by internal Microsoft IT teams.
  • MSCOM - The server for development done in the Microsoft.com organization.
  • PG Consolidated - A collection of "smaller" (generally less than 100 people) teams from around the company.
  • MSN - The server used by a collection of MSN teams (although the entire MSN org has not adopted TFS yet).
  • EVAL OFFICE - Office is currently in the process of evaluating TFS for use in developing their next release.  We set up this server for them to experiment with to help them get familiar with TFS, decide what customizations they will need to make, etc.
  • SQL - The server being used by the SQL Server team.  They are only using TFS for work item tracking right now.
  • EVAL SPCore - This is a general eval server for teams that want to try out TFS.  "SPCore" derives from the name of our internal adoption team - the Solution Provider team.  They help teams with the process of adopting TFS.
  • EVAL * - More Evaluation servers.
  • Pre-production environment - A set of virtual machines that we can use to load images of production servers to test various maintenance and upgrade procedures without putting production machines in jeopardy.
  • Monitoring - We have a set of monitoring servers that perform various tasks including measuring availability and performance of other servers and mirroring checkins and work item changes between TFS and the internal team development tools we are gradually migrating away from.
  • Extranet - a server set up for external partners who collaborate with Microsoft to use.

I've attached the Visio diagram to this post if you'd like a much higher quality rendering of the image.

Published Tuesday, August 22, 2006 11:29 AM by bharry

Attachment(s): Dogfood Topology.vsd

Comment Notification

If you would like to receive an email when updates are made to this post, please register here

Subscribe to this post's comments using RSS

Comments

# Microsoft's Team Foundation Server Deployment Topology

Tuesday, August 22, 2006 12:35 PM by Rob Caron
See Brian Harry's blog to understand the scope of our Team Foundation Server deployment. The image on...

# Team Foundation Server Use at Microsoft

Tuesday, August 22, 2006 12:59 PM by .net DEvHammer
bharry talks about how we use Team Foundation Server at Microsoft, and even provides a nice pictorial...

# TFS Internal Dogfooding

Tuesday, August 22, 2006 4:36 PM by Ken's Place
Brian Harry blogs about our current TFS Dogfood topology at Microsoft. 
"Given how much I talk about...

#

Wednesday, August 23, 2006 3:54 AM by MSGuayaquil

# Microsoft's Team Foundation Server Deployment Topology

Wednesday, August 23, 2006 4:01 AM by 寝ても覚めても.NET(?)
bharry's WebLog からMicrosoftがTFSのDogfoodを行ったときのサーバー構成などに関する覚え書き?に近いもののようです。Microsoft's Team Foundation...

# re: Microsoft's Team Foundation Server Deployment Topology

Thursday, August 24, 2006 8:31 AM by bharry
A couple people have noticed that some of our instance configurations have 3 servers in the little box.  They have both the AT and DT and another server - for example tkbgitvstfas01.

They further observe that none of our public documentation describes any scenario where by this can be done.  Good observations :)

Those "extra" servers on the diagrams are SQL2005 Analysis Services Servers.  Basically, they are the cube process portion of the TFS data warehouse.  For a couple of our largest instances, we have off-loaded this work onto a third server to reduce the impact of reporting on the operational stores.

This configuration is NOT supported in V1.  We made a couple of fixes in the coming TFS SP1 that enable this.  Once TFS SP1 ships, we'll produce a Knowledge Base article that explains how to do this.

A word of warning though - it still won't be a configuration for the faint of heart.  We don't have tools to automatically set it up and it requires openning up SQL Server Management Studio and modifying some data.  I would only recommend this configuration for people who have servers with VERY high levels of data/load on them.

In a future version we will make this and other topology configurations more main stream.

Brian

# Martin's got a confession

Friday, January 12, 2007 8:22 AM by Jeff Beehler's Blog

I'm not sure if size matters, but if it does, Martin's laying claim to the World's Smallest TFS as compared

# ¿Cómo Estructuró Microsoft su Deployment Interno de Team Foundation Server?

Saturday, May 05, 2007 10:19 AM by Julio Casal

Ok, Ok, toda empresa de desarrollo seria debería ya estar al menos pensando en implementar Team Foundation

# Hello

Wednesday, June 18, 2008 5:18 AM by Grant Holliday's blog

You might remember me from such blogs as OzGrant.com , mailing lists such as OzTFS.com , the Team System

# Team Foundation Server Proxy blog

Thursday, June 26, 2008 9:41 PM by Grant Holliday's blog

Before I joined MSFT I hadn’t had any exposure to Team Foundation Server Proxy. As you know from Brian’s

Leave a Comment

(required) 
required 
(required) 
 
Page view tracker