This week was another nail biter in the Virtual Design Master competition. Challenge 2 required us to save the moon while using someone else’s design plus a few constraints: must fit in 21U, have to use the same vendors (but can use different product lines) as the provided design used, and the big one, the moon base only has IPv6 networking. I understand IPv6 but certainly haven’t designed an IPv6-only network, so this was pretty scary and very time consuming for the research.
There were a lot of great designs presented by the VDM competitors. Three of us had to work off of Daemon Behr’s and six of us had to work off of my design from the previous challenge. It was fun to see how other people managed the same base project and morphed it into a project that had their fingerprints on it. Watch the results show and check out the designs (here’s mine). During the design and the judging, I learned a few things in no particular order:
- Vendors are inconsistent at stating their IPv6 support stance for their products. You might have to track down an SE on twitter or in person to get the scoop, or you might get some vague information or conflicting information, or even no information at all. I made the assumption that no mention == no support, which is tolerable for design but a day or two in the lab could prove otherwise.
- When you do find support statements, do not be surprised when there is no support for IPv6-only networking. vSphere itself supports IPv6 but doesn’t support IPv6 iSCSI with 2 of 3 initiator types; VSAN and Log Insight have zero IPv6 support; vCO supports it. Only one out of three, and it needs to be 100% for IPv6-only to work.
- Everything you choose is a risk. A or B? Both risks. The Giant will kill you anyway.
- Well, the judge is not going to kill you. But regardless of which design decision you took, be prepared to defend it. Neither A or B are wrong, so don’t think you’re “safe” because you choose the “right” one.
- If budget isn’t an issue, do not undersize your solution. I used 12U out of 21U (I did have one good reason – with no redundant rack, one unit overheating could damage another touching it – but that’s a stretch) and said “…if the design scales up.” This is a last ditch effort at saving humanity. There is no next ditch effort. Get 21U of equipment, or have a better reason than I did for not using it all.
- MB is not the same as GB. Whoops.
- The competitors have a great sense of co-opetition. While we all did our independent research,
- I better learn about NetApp snapshots, or Josh Odgers may ask me a third time. That would not be good!
Again, I was pleasantly surprised to find that I survived to the next round. On top of that, the judges said that my design was impressive! I’m still stunned, but also very proud. I must again thank those who reviewed my proposal and offered advice. I definitely wouldn’t have made it without your help.
Challenge 3 actually terrifies me. Between now and Tuesday at midnight, I have to learn OpenStack, build a design based on it, then actually lab it up and provide video of the proof of concept! I always wanted to learn about OpenStack… Wish me and the other competitors luck!