Microsoft Tech Ed 2013 with the EMC MSpecialists

Well, I am back from my first Microsoft Tech Ed (held this year in New Orleans). There were a lot of interesting items there, not the least of which were the amazing deals on the Surface RT and Pro. And yes, I did avail myself of one of each.

My focus is generally Microsoft Exchange and the sessions on Wave 15's Exchange 2013 were at times validating, other times interesting, and yet other times puzzling.

Items that were validating:

• the Exchange Sizing overview for E2k13 was great. At EMC, we were waiting with baited breath for the spreadsheet and non-Kool Aid performance information to come out. While I suspect that the BDM reductions will need to be validated in the lab, the mechanics of it (tweaks for checkpointing, lengthening the amount of time forBDM completion) makes sense.

Items that were interesting:

• Managed Availability for Exchange 2013. This comes as a direct result of Microsoft changing its policy for support by making Exchange Team developers have responsibility for On Call and support. By making them "feel the pain" from the customer side,  they were obviously incentivized to automate recovery. Leveraging probes and simulated transactions, E2K13 essentially brings out a series of increasing size "hammers" (i.e. service restarts) to automatically effect recovery, re-seed, instead of just activating a DAG copy.
• Dog Food. Microsoft pulled the covers back from their internal environment and were upfront with the fact that the released code is being executed internally first - and at scale. Why is this important? Along with Exchange 2013, there is a new quarterly cadence of releases. No more CU - now there is a NEW RELEASE of Exchange every three months, with support extending for the prior version another 3 months. Now is as good a time as any to mention the puzzling items....

Items that were puzzling:

• Exchange DAG deployments are relatively complex (compared to, say, 2003 days). Smaller shops find the architecture to be somewhat daunting. While this has been partially addressed in the flattening of the Exchange roles, there was still the option of implementing a single multi-role node for smaller deployments. If you are uninstalling Exchange every 3 to 6 months to stay on a supported version of code, this pretty much forces you into DAG deployment.
• I seriously recommend that people take a look at the new caching algorithms for E2k13! Memory was a big part of what changed in the new release, and there was a nugget dropped in the Tips and Tricks session (not posted yet - will port link when available). Big thing to know: IF YOU ADDED EXCHANGE DATABASES, THE CACHE WILL NOT BE OPTIMIZED UNTIL THE STORE HAS BEEN RESTARTED! The algorithm is optimized at startup. So things will kind of stink until things are restarted on the node.

There were some REALLY neat things announced for Windows Server 2012 and Hyper V coming in the Windows Server 2012R2 - I will be writing about them in an upcoming post.

What really blows me away about 2012R2? Holy cow - they are not going 3 to 5 years any more! The cadence is fast and furious (at least for now)!

All of the sessions are up on Microsoft's Channel 9. I strongly urge that you take a look at the videos and decks - with this new pace its more important than ever to stay current!

Lastly, one other thing happened at Tech Ed. Sam Marraccini (from EMC Inside the Partnership) got a video of Hacksaw doing the ol' Humpty Dance on Bourbon Street. What happens in New Orleans apparently does not STAY in New Orleans!