An Inside View: The Road From Beta to RTM

Hello Community:

I have not blogged in a while, so I thought I would provide some insights on what the team has been busy with since releasing Beta 3 in June. First, we have released a number of updates to NIS through Microsoft Update, our Beta 3 tools and SDK and also updates on how to use System Center to monitor the health of your TMG deployment. Next we have been receiving a lot of great feedback through not only our Beta and TAP programs, but also through various telemetry reports you have been submitting. I cannot state enough how valuable these are to us in improving the quality of the system, protocols and signatures. We have also been examining Dr. Watson reports from both beta customers and our large ISA community to fix the most common issues reported by you through the customer feedback program. Believe it or not, all the data you send to us is examined and analyzed very seriously and we use to improve the product constantly based on the issues you run into. In addition, in the area of collecting data, we are working with the MRS team on improving the categorization and accuracy of the URL filtering in the TMG Beta. This is a work in progress and it will continuously improve as we get closer to our final release. The beauty of the system is you do not need to update TMG to inherit the improvement, it will happen naturally by using the service. We will be releasing a telemetry pack in the next week or so (please watch for it) that will help us to get the data we need to improve the categorization specific to your businesses.

A second area we are focusing on is our performance, stress and reliability testing. We have heard loud and clear from the community that you want capacity planning data. We are working on that as I write this and we will have the data available to you as soon as we have completed our testing and can document the results in a consumable form. We are really excited about our initial performance results and the capability of the system, we are sure you will be as well when we can publish the data. The testing and the measurements do take time. For example, running the system under stress for weeks and watching for potential memory leaks takes a significant amount of calendar time. A single bug could cause us to reboot and start the cycle all over again, so this is one of the reasons we take our time between a beta and the release candidate build.

Next, we have been updating our management interfaces based on feedback and usability tests. One example that I personally got involved with was the ISP Redundancy Wizard which we found was a little confusing in the beta. We have re-worked this wizard to make it more intuitive and easy to use. We are also updating inconsistencies between wizards and property pages – thanks to some of the MVPs out there who have been catching these and providing the detailed feedback! Finally, in this area, we have begun our initial localization of the product into multiple languages and testing in multiple languages to ensure that everything displays, aligns and can be viewed properly.

Of course we are performing an additional round of security reviews, security testing, bug reviews and threat analysis. This all takes time and we review each potential issue very carefully to ensure we have the highest possible quality product when we release. Again, this is where the bug reports and Dr. Watson data helps us to track down undiscovered issues proactively and fix them. We then need to apply all of our various configurations, workgroup, domain joined, small arrays, large arrays, Windows Server 2008 and now Windows Server 2008 R2 platform testing for our entire set of automated and manual tests. Let’s not forget ensuring the system is ready for unattended setup and delivery as a VHD! It is quite extensive and exhaustive when you look at the scope of the effort.

Last, but not least, we are investing heavily in analyzing various troubleshooting and what we call "supportability" scenarios. What I mean by supportability is how easy is it for a customer to discover an issue and resolve the issue easily or with minimal time and effort. We have placed a very specific focus on this aspect in this release as TMG is not just another release of ISA, it is a revolution in functionality and capability. We want to match that revolution with world class reliability, ease in management and overall simplicity of operations.

In the end, we are very busy this summer getting the product ready for release. We are joint testing with all the latest versions of our partner team components such as Exchange 2010, Forefront Codename "Stirling", UAG, SQL Server 2008 and of course Windows Server 2008 R2. We will have a final RC (release candidate) in the fall – so if you have not downloaded and tried out Beta 3, please do so and help us find the last few bugs specific to your environment, configuration or network!

David B. Cross

Product Unit Manager

Forefront TMG