VMware Memories, part 1

[#1]

Sitting here on the eve of my 7th VMworld conference (I missed the first one in 2004, but have been fortunate enough to attend all of the others), I think back on VMware journey up to this point and the people who have influenced me along the way. I’m a fairly inconspicuous guy, so I don’t get recognized a lot, even though I’ve been around the VMware scene for quite some time.

In the beginning, I used VMware Workstation (2.x) when I was a software developer. None of us ever wrote buggy code and had to reload Windows (3.1!) onto our development machines, but Workstation was there as a testing environment, just in case. What a timesaver! To be honest, my history with VMs goes back before that with RealPC, VirtualPC, SoftPC, and the other emulators that ran x86 machines on Macintoshes (it worked, but was NOWHERE near fast). I always wanted something like that when I was developing software on my Mac – but that never happened. At least, not during my developer days: vSphere 5 allows it if all of the prerequisites are met!

In 2002, I continued my use of Workstation (3.x) during project planning and new software testing when I moved from the app/dev side of the house to infrastructure. In those days, I did a lot of directory services work, and testing integration plugins or new Netware client builds/patches without nuking my production machine's configuration was the primary use case. Moving beyond my own machine, we licensed GSX Server for Windows in 2004 and collapsed two racks of lab hardware onto a pair of servers. Our lab room immediately became much quieter, cooler, and more available. Those who understand the ‘waterfall’ model of lab gear acquisition get that labs are regularly built from the oldest junk in an organization -- and hardware support is not even within the realm of possibility.

After such great success in our lab, and with the 2.0 release of VMware ESX Server, we decided to take a look at what that could buy us beyond our GSX Server – we’d heard good things about less overhead and higher densities. Back in those days, you were required to engage VMware resources in order to even try out ESX Server. Enter Melissa Ercoli and Kenon Owens – these people were my first experience with corporate VMware, and helped me get started. I believe Melissa got married and is still with VMware while Kenon has since moved on to work at Microsoft. Both of these individuals invested the time with me to ensure that I understood this new ESX Server thing by answering all of my questions, both technical and licensing-related. For that, I am very grateful.

In late 2004, armed with an understanding of the ESX Server architecture, and a trial implementation in the lab, I submitted a proposal to our management to stand up a POC environment in production – 2 SAN-attached ESX Server 2.5 hosts. I was immediately met with the “all of our eggs in one basket” argument. Fortunately for me, VirtualCenter 1.0 had been released and I was on top of the amazing VMotion (notice the capital "V"!) technology enabled therein. My POC environment was kind of sad – we basically took two of our standard x86 servers (HP DL360 G3), bumped the RAM a bit, and shoved an FC HBA into it. There was pretty much no redundancy (single power supply, single SAN path, single NIC port for Service Console, one for VMotion, and two for VM traffic), but the thing worked. My first demo of VMotion blew everyone’s mind: “there’s NO WAY that just worked!”

In August of 2004, a hardware failure on a production server expedited the migration of our VMware POC into production. We were presented with the option of blowing away the POC environment in order to repurpose its hardware as a replacement for the failed box, or repurposing a VM currently running on that environment. The path of least resistance, and quicker TTR, was obvious, and we became production ESX Server and VirtualCenter users. From there, the thing spread like wildfire, just like the stories we hear today -- you know, the Legend of VMware.

This stuff was getting to be fairly complicated and I had invested a lot of time into understanding how the pieces fit together. I registered on the VMware Community forums in October of 2004 in order to get answers to my more complicated questions and to share my experiences with others. First off, I was amazed at the sense of community -- people helping people without being compensated beyond the satisfaction of sharing knowledge. This was a group I immediately liked quite a bit. It was there that I ran across the VMware Core Customer program, which ultimately led me to Christine Holland. Chris was interested in my environment and how we were using VMware’s technologies in both production and testing, so she introduced me to the PR people and they quoted me as part of the ESX Server 2.5 release. We had a case study published in early 2005 and a related article as well.

From there, Charles Babcock (Network Computing) and Jennifer Mears (Network World) talked to me a bit about the ESX 3.0 and VirtualCenter 2.0 launch, and the much-anticipated HA and DRS features. I had some coverage there, including a quickly-snapped photo of me at VMworld. I’ve got to say it was pretty cool to get a magazine with my (goofy) picture in it!

To be continued...