Considerations for Hyper convergence - CTO Coffee CHAT
Transcript
I'm doing so this is a little bit therapeutic as it as it is hopefully beneficial for you guys on periscope first I hope you'll notice that i have my emt electrocution errs electrocution errs official dodgeball or vouch balls turned on from vmworld 2014 I think I think I played dodgeball in 2014 I'm doing the Sydney a/v bug on i think the 23rd of februari and the topic is went to select type of convergence so as a preamble to that I thought I'd talk about some of
the decision-making process I'm going through currently in my current organization one of the things that we have to decide about decide upon is well not to invest in ASAP appliance as it'd be hotter appliance or to do what they call TDI which is basically doesn't matter what it stands for is just basically build your own and bring your own solution to the forefront and this is one of the ideas I want to capture and doing this talk on when it's that type of convergence a lot
of times we spent a lot of the time or a lot of our time focusing on technical requirements I almost definitely go over a lot of the technical requirements when it comes to hyper convergence but one of the topics that I want to touch on is from somewhat of a business or business driver perspective or a non technical piece which mainly is around support the things that that a compass or influence the support model in the SI p hana world SI p has traditionally pushed the
appliance not primarily for technical reasons over the past couple of years originally the technical reason was that most people infrastructures weren't prepared for hana hana from a high level should include 10 gigabyte 10 gigabit ethernet a fairly fast or I'm constraints and so not necessarily fair science and butts and that wasn't can there was no contention and a sand fabric that could support the low latency required for essay p Hana and then obviously from a database perspective from hanas perspective they needed a certain intel chipset
until CPU and chipset so in order to ensure this they came up with the appliance model benefit which is the thing I want to talk about mainly is the ability to pick up the phone say that I'm having performance issues with my essay p Hana environment and get anti end support you can call as AP directly and they'll support the appliance all the way up through the database in the application stack that is I think indicative of the hyper converged argument so EMC yesterday I think
it was announced VX rail I get thanks for the EMC elect privilege of getting briefed on it a week early uh it is a pretty nifty solution so if you think about hyper converged in general you know you get commodity hardware that's tightly integrated with the software stack the hardware itself is commodity at this point it doesn't matter if it's super micro dell a emc specific spec the point is is that you have centralized support to call and one experience for mission-critical or in this case
not mission critical workloads I I don't know if I'd necessarily put VX rail on I don't know if I put s ap or a mission-critical app on a VX rail not necessarily the the use case the use case is probably where you would have this node in a or several of these nose and a remote office some type of edge solution where you don't have IT staff or you might have IT staff but the point is you don't have the death of experience needed to support
a complete stack so it's critical in these environments that you can pick up a phone and call one person in this case EMC doesn't matter if it's MC new tannic simplicity scale the list goes on and on but basically one resource to call to support the whole stack so as we expand out our footprint uh you have you're deploying hundreds or dozens of these things the convenience of the support model makes sense the technology you know what doesn't matter if you deploy a hyper converge solution
a direct attached storage solution or some of some other technically feasible solution probably doesn't not from a technology perspective not and those small or the increments not when you're talking one or two nodes it is mainly the the value is not in the hardware or the technical solution itself is in the support model that's y VX rail is an important solution that's why companies are still considering SI p hana appliances even though most of our infrastructures hopefully if you're considering SI p hana that's why most
of our infrastructure is now support the minimum requirements for a memory database some of the things i'll be talking about and i'm currently putting together for my FEMA presentation for your folks in Sydney you should be worse that's what ten ten eleven o'clock so you guys should most definitely be up looking at these videos getting prepared getting excited for me to make the the trip uh I don't know if you can call it the other
