RedefinePossible

-> come close?

not yet! But there are more data services now that folks can try.

EMC has not yet had to pay $1m - keep trying IT guys get the dough Goulden dared you at @emcworld come on

<marketing spin=off> if you could do it then we wouldn't be offering a million dollars.

@jburton Hopefully most people realised that. It's a fundamental transaction flow in the platform, which of course we know isn't the same for Pure.

separate.

XIO is an all flash array - tightly coupled scale out cluster.

ViPR's storage system block / object / hdfs against commodity

So what is the upgrade path for users? HP's all flash array, for instdance, is built into the same architecture as its disk arrays.

ViPR controller will "cloak" both XtremIO and ViPR's storage system and present a consistent set of interfaces northbound to things like VMW and Openstack.

ViPR is essentially an orchestration/automation layer - mostly tech from the iWave acquisition

From what ?

@chrismevans VipR is 2 things - controller (orchestration layer) ... and new storage system (block, file, object) on commodity

No upgrade/migration path from VNX/VMAX to XtremIO. Migration task.

Then what holds EMC customers tdo EMC when they start migrating Tier 1 to all flash arrays?

if you build for disk, you compromise. Only guys who don't have a legit all flash array say otherwise. We have both.

@jburton Yes, agree, but at the data layer, none of the components interact (e.g. migrate from EMC to NTAP)

@chrismevans expect the migration layer to come as a ViPR service.

One option is to virutalize with VPLEX, no? but again you're adding yet another layer to be able to move data from one array to another. Starts getting complicated and expensive having to purchase/license everything

@silvestri_73 The problem with replication is almost all of it (natively) is dependent on the hardware architecture (replicating blocks/tracks/objects), which is where it gets tough, hence abstraction.

@chrismevans agreed, but replication at it's heart is within or external to the box. seems odd that the internal replication function of XtremIO is native, whereas the external replication tool is an add on.

@silvestri_73 Could be described as a failure in design, or perhaps a whiteboard feature not yet implemented. Original AFA's weren't good on features! :-)

yes that was me

Pictorial evidence abound!

Right now we're doing pretty well with VCE and VSPEX. IDC dont count 50% of our VSPEX revenue so we're actually doing better there than most folks thing.

We have some wild stuff under wraps that will form the basis for the next generation.

at scale CI systems will be disaggregated pools but managed as a single system.

VCE has been crushing the marketplace in terms of uptake

I prefer to call it an "ocean" but that is not taking hold..lakes sound so small and not dynamic

Is this the "nicer" version of Data Landfill?

Lake is not the right word.

Better than Data Pond

@JeffFrick oceans have currents; streams too; lakes are stagant that is #DWBI of old

Agree. I don't like "data lake" but it seems to be the term people are using.

Think of the great lakes...

@BertLatamore I didn't like "Big Data" but there we are, we're stuck with them unless we invent them!

"data leak" :)

I think for many companies that can't analyze their data today, the dormant "lake" is a good image

And better than data teapot or data bathtub

got to this one late.

have heard this thing called a "Data Landing Zone" as well

folks want to dump data into HDFS and not go through the data warehouse ETL hell.

dump it, map reduce it, analyze it. More stuff in one place... better information. Retailers, banks are leading edge

@jburton are you happy with your decision 4yrs ago with the them Cloud Meets Big Data? Whats met your expectation and what didn't?

Big Data Jailhouse?

And who says a lake is small? The Caspian Sea is considered the largest lake in the world! ;-)

@jburton sounds like aircraft carrier on the "ocean" - i'm not giving up on my meme of data ocean :-)

I don't think a terrestrial oriented metaphor can do data justice.

that seems to be a good strategy as #VPLEX and #Recoverpoint are becoming easier to consume as Software too

true... but can a virtual vplex/RPA scale to the performance of an XtremeIO?

we have Moore's law in our favor :)

and if shared compute will not do it, the appliance model will be the answer. It is all about scale.

agreed. EMC's risk is unhappy customers that try to deploy virtual VPLEX/RPA and get bad performance. As VMWare discovered by recently reducing the supported VSAN controllers

just because you can do something, doesn't mean you should

agreed 100%. Every design has its implications. The use case defines the boundaries.

both supported today... in fact many customers now purchase XIO with VPLEX.

got some stuff coming to make it much slicker in the not too distant.

The chip guys keep warning thatd we are reaching the end of Moore's Law. You can only cxram so many circuits onto a silicon chip.

I suspect that is why we are seeing more specialized chips/systems -- a way to increase performance for specific loads.