Highlights from Fall Exchange Connections 2008

Posted November 14th, 2008 by Jonah Paransky

F08_Exch_Mid_REV Today’s post is a wrap up of the Exchange Connections 2008 conference where StackSafe exhibited this week. I had a chance to speak with many interesting exhibitors and attend some great sessions at the Fall Exchange Connections 2008 show, but I thought three of the sessions covered content of particular relevance to the readers of IT’s About Uptime.

Migration to Exchange 2007: the Front-End by Robert Dawson at HP

Robert provided a detailed overview of the process involved in the front-end migration of Exchange 2003 to Exchange 2007. He covered a variety of topics including:

  • Prerequisites of coexistence
  • Installing Exchange 2007
  • Making the switch
  • Links/References

Listening to the overview, Robert made a number of points that provide valuable lessons to a wider variety of upgrade projects.

  1. Plan – If you fail to plan, you plan to fail
  2. Gather information. You need to understand the environment before beginning the migration process. Robert particularly raised the issue that often teams overlook 3rd party applications that may rely upon the existing exchange infrastructure.
  3. Get the right people together for the migration. Exchange environments are often large, encompassing Active Directory servers, messaging components, security teams, network teams, etc. It is imperative to bring everyone together to properly prepare for the migration

Building an Exchange Test Environment in a Hurry by Michael B. Smith of The Essential Exchange

Michael covered the steps required to build a “quickie” test version of Microsoft Exchange in a virtual environment.

Michael made two points that relate to pulling applications into virtualized environments that seemed of particular interest:

  • Beware of memory contention – some applications are memory hogs. Even in a hypervisor-based virtual environment, it is critical to properly size the server so that the performance of memory intensive workloads are appropriately maintained
  • Beware of I/O contention – some applications have significant I/O requirements. Often the limiting factor in the performance of virtualized workloads is I/O, so keep a close eye.

I wish I’d known: Exchange 2007 Upgrade Lessons from the Field by Jim McBee of Ithicos Solutions and author of the Mostly Exchange Web Log

Jim covered lessons he has learned in the field performing multiple exchange transitions. A number of these lessons were universal as well, and worthy of mention:

  • In larger organizations, don’t underestimate the amount of time and resources that will be required to prepare the Active Directory (AD) infrastructure. With tasks ranging from removing (or moving) legacy Server 2000 controllers to updated schema, Active Directory prep can be a significant undertaking. Jim specifically mentioned an example where the preparation required a 6 week approval cycle for a single AD change.
  • Make sure to find all the existing applications that may rely upon the Exchange environment prior to beginning the transition project. Jim used an example of a customer that had 1000 applications that relied upon the Exchange infrastructure, each of which needed to be analyzed and tested prior to the transition.

Where does testing play in all of this?

All the sessions provided great detailed information. One thing that felt missing (perhaps due to our own focus on the topic) was a deeper look at actual testing activities required to insure a successful transition. A number of unanswered questions popped into my mind during all the sessions…

  • How do you test to make sure the required Active Directory changes will not negatively impact production prior to release?
  • How do you test the impact of the migration on related applications, before the infrastructure is deployed into production?
  • How do you make sure that the proposed server roles are properly designed, before the infrastructure in active in the environment?
  • How do you ensure a smooth migration, without production problems or disruption of mission critical messaging services?

I would like to hear your thoughts on any or all of these questions.

Popularity: 4% [?]

Filed Under: IT Operations, Testing, Virtualization


Leave a Comment