Itzikr's Blog

EMC World 2013 is behind us and i thought i will be a good idea to describe my experiences from the EMC puzzle piece I’m working for, aka, XtremIO.

The way to Vegas..

im not a big fan of flying to the west coast, it’s around 17 hours of flying and with the connection time, we are talking about 19-20 hours, however, on the way in, a colleague of mine, fainted out, luckily that with the help of a DR that was on board and to some extent with my ex military skills (Combat Medic), an IV line was opened.

unluckily for me, a bad cough turned into a serious “loosing my voice” since i had to cover for the guy on the cold airplane floor, good fun!

The Hands – On Labs

the HOL’s as we call them are one of the conference crown jewels, it’s a Hugh amount of work and sleepless night for the vLab team that is responsible for it, see a separate post here to understand the dedication required:

http://www.emc.com/emc-plus/article.htm#/blog-alan-mcgrath-senior-manager-of-vlab-content-engineering-for-emc-has-been-working-day-and-night-to-make-the-hands

but for the XtremIO team, it was an extra special, this year, we put couple of X-Bricks to carry the load of some super heavy IO vLabs, did it work perfectly? NO, we had some gotchas that we had to find workarounds while running so that’s always fun, interestingly enough, the majority of the bugs were to do with vSphere 5.1 and 5.1 Update 1 fixes a lot of them, see a separate post here:

https://volumes.blog/2013/04/26/heads-up-using-xtremio-and-vsphere-5-1-update-1/

unfortunately, the update was never applied before the show so we had to find some other workarounds, always a good fun!

The FLASH Keynote: 

Zahid was doing a great job presenting the FLASH team (Both the XtremSW and XtremIO), the fact that he brought a customer on stage (Sandy Bryce from Bailie Gifford) just gave more emphasis to the point that XtremIO, despite being in a Direct Availability (limited) Stage, kick Ass!

see the full keynote here:

apart from the FLASH Keynote, XtremIO was mentioned in other general keynotes as well,
Pat Gelsigner talked about it in the context of the best array for VDI, see the full keynote here: (around the 26th minutes)

we also had an EMC TV Interview – Wednesday: Backstage Session with Dan Cobb, CTO, FLASH Products & Technologies, EMC & Robin Ren, CTO, XTREMIO Division, EMC:

Trey Leyton (VCE CTO) was interviewed about the new VBLOCK and gave some hints about an XtremIO based VBLOCK to come..

we had two sessions this year, repeated twice and presented by our VP of Marketing, Josh Goldstein, the main one describing XtremIO, as you can see from the picture above, was standing only, i cannot imagine what will happen once we actually GA, the audience interest is staggering.

The 2nd session was about the use cases, what was so uniue about it, is that we let REAL customers who are using the product to describe their findings, honestly, i cannot asked for better speakers

The first customer to speak was CMA, they described how they are running their Oracle RAC instances on XtremIO

they were able to cut down their latency to < 1ms and drive crazy amount of bandwidth to the array which they were never able to do before!

The second customer was Sandy Bryce from Bailie Gifford, they are using XtremIO to take their production Oracle DB instance and to clone it many many times while still having the same performance of the source DB and while still utilizing the capacity of the source lun only, talking about eating the cake while leaving it full..

Sandy Bryce from Bailie Gifford presenting

the 3rd customer to come on board was Boston Scientific, they are using XtremIO for their VMware VIEW VDI environment, there are couple of things that were unique about them, the first one is that they are using persistent “thick” VM’s, as you may be aware, while linked clones can be lighter on arrays with large cache, they DO possess a large administration burden, see jim sanzone talks about it here:

the 2nd thing that was unique about them, is the dedupe ratio they got from using XtremIO, prior to this they were already running with an array that is supporting a (post-process) deduplicaiton but our dedupe ratios were much much higher..why?

because we are doing it on the entire X-Brick or the Cluster or the global level!

boston scientific presenting their VDI project using XtremIO

The Booth

the booth was packed with interest, customers coming and asking how does our architecture differ from what’s already in the market and after they get the difference, you could see that it “clicks in”

Customers Meetings

probably the best part of the event for me was the customers meetings, thanks to my manager, my diary was packed with back to back meetings so i couldn’t attend sessions but boy, i could speak..

few interesting observations here:

customers do ask what is the best way to start with XtremIO, what use cases will fit, does using XtremIO means i don’t need my VNX / VMAX for anything? the answer for this is no, XtremIO is really good for specific use cases but for other, other products in the portfolio may provide a better fit, the advantage of working for EMC is that we don’t have one product so we don’t adjust the customer needs to the product 🙂 this philosophy is in the root of EMC and will be for the years to come, what will happen in five years? i do not know but we are lucky enough to be able to “cannibalize” our own business than lettings other do it for us.

the other observation i found interesting is that some customers thinks that XtremIO is hugely expensive, it is not! i can’t wait for the GA to share more.

Until next time..