VMUG London January 2015
Despite having attended the annual VMware-hosted London vForums a few times this was my first VMware User Group meeting.
The all-day event, held at the London Chamber of Commerce and Industry, featured a number of talks from, and discussions with, both suppliers and the VMware user community.
A Giant Mountain of Awesomeness
Following the introduction to the day, the first presentation was from PernixData and CloudHelix on using their FVP software to improve storage performance- at a very basic level, it uses flash drives and memory located on the Virtual Host layer to deliver more IOPS, allowing the architect to scale the SAN for average performance rather than the peak, and avoid the issue of stranded storage- capacity that can’t be reached because the system has “run out” of IOPS. I’ve seen PernixData before, but this was still an interesting presentation- bringing in scenes from Spaceballs, The Flash, and Flash Gordon whilst talking about a 1.6 million IOPS “Giant Mountain of Awesomeness”.
Flash Gordon references in a presentation are always a bonus. #LonVMUG pic.twitter.com/7u8Q0secly— Chris Bradshaw (@aldershotchris) January 22, 2015
Riding the Lightning
Next up were the vFactor Lightning Talks- these were 5 short presentations delivered by community members, rather than vendors, which focussed on using VMware products (rather than buying them). I really enjoyed these talks, they were full of useful nuggets of information and it’s always nice to see where others have encountered the same issues as you (or the issues you’re about to get!) and how they’ve overcome them. Philip Coakes (@CoakesPhilip) gave some good advice on saving money and spending wisely, summed up by his comment of “Spend your company’s money as if it’s your own”. Alec Dunn (@LegoYoda) also had some important things to remember:
"Once it's working, back it up" - good advice for any new service from @LegoYoda #LonVMUG— Chris Bradshaw (@aldershotchris) January 22, 2015
Swimming with Whales
After those presentations I attended a session on where VMware interacts with Docker presented by Andy Jenkins (@stonestokie) and Robbie Jerrom (@robbiej) from VMware. Everyone had heard of it, a few were using it in Dev/Test but no-one in the audience had production loads on the platform. It’s interesting to see the VMware take on Docker- it has many functionality similarities to their virtualisation platform, and is something that VMware (along with Microsoft as shown in this timely blog post from Ed Baker) are keen to show can be positioned well into a traditional virtualisation stack. I can see the benefits here, the ability to rapidly spin up potentially massive numbers of the Docker applications and the hybrid cloud hosting resilience are enough to convince me not to uninstall ESX just yet.
Is there a DR in the house?
Next on the block was a presentation from Unitrends on Datacentre failover, based around their ReliableDR product. Ian Jones highlighted the two biggest problems with datacentre failover- Failback and Documentation- and demonstrated how their product automates the process, even using existing Array/Software based replication between sites. The automatic testing (and documentation of said testing) appeals to me, but led me to wonder how easy it would be to get complacent that everything was working- do we need an “automated DR tester” tester? :)
“Bringing the SDDC to life- a real world deployment”: A detailed talk from Phil Monk and Jon Kemp (both VMware PSO) and Michael Poore (Xtravirt) covering a lot of the design work in deploying a Software Defined DataCentre. This talk was much more project orientated to start with, diving into how the design process worked before looking at the conceptual, logical, and physical designs and some of the decisions behind them. An interesting mix of process and technical details.
My sessions were rounded off with a presentation from Valentin Bondzio (we were told “Cloud is in the other room, and that’s a fad!”). This was by far the most technical talk of the day, and thankfully the distance between the CPU core a process is running on and the cache it’s data is sitting in is not a day-to-day concern for me. Despite the technical depth of NUMA, memory access latency, Ready states and locality, the clear presentation style meant I came away from this thinking I had definitely learnt something I hadn’t known that morning.
All in all a very good day- a wide variety of topics without a heavy sales pitch- followed by some vBeers (thanks to 10ZiG for the sponsorship). My first VMUG, but hopefully not my last.
All #lonvmug slides (including Lightning Talks) have now been uploaded to http://t.co/FZaJKUX47z. Thanks to all who contributed yesterday.— Alaric Davies (@alaricdavies) January 23, 2015