drivescale_logo-100x27DriveScale – another company presenting at Tech Field Day 12 that I know very little about – consider this post a homework assignment on my part – to at least learn a little bit about the company, the problems they are trying to solve, and the products and services offered.  Just as the last company I had a look at, StorageOS, DriveScale is relatively young in the IT vendor space.  Not to say that is a bad thing – normally startups execute quickly and solve real world issues that exist today.  DriveScale has been around since 2013 but just come out of stealth in May this year – so naturally, this is their first appearance at a field day event.  Before we get into what DriveScale does and how their technology works we should take a look at something that peaked my interest right off the hop- and that’s the founders – In order to best understand this let me list each founder with some highly summarized bulleted accomplishments – I think you will be a little impressed.

Satya Nishtala

  • Holds 31 patents to his name in core datacenter areas – 31, yes, 31!
  • Technology Fellow at Nuova (Eventually acquired by Cisco and baked into UCS/Nexus platform)

Tom Lyon

  • Founder of Nuova (Eventually acquired by Cisco and baked into UCS/Nexus platform)
  • Employee #8 at Sun Microsystems – think about that for a minute – Sun Microsystems, Employee #8

Duane Northcutt

  • conceived of and led development of the Sun Ray desktop!
  • Held CTO/Vice President positions at Technicolor, Trident, and Silicon Image

The list goes on and on for these guys but man those are some hefty accomplishments to have at the helm of one company for sure.  Anyways, what they have done is not as important as what they are currently doing now, so let’s have a look at that.

Disclaimer: As a Tech Field Day 12 delegate all of my flight, travel, accommodations, eats, and drinks are paid for. However I did not receive any compensation nor am I required to write anything in regards to the event or the presenting companies. All that said, this is done at my own discretion.

Smarter Scale-Out

DriveScale’s whole solution is based around being a smarter scale-out solution – offering a rack scale architecture which includes both hardware and software to bring “advantages of proprietary scale-up infrastructure environments to the commodity of the scale-out world”  <- When I read this I kind of though, huh, I don’t get it – It sounds good, but I really don’t know what it means.  This is mostly due to the fact that they really target Hadoop and big data environments, something I’m not well versed on at all!   I’m sure we will all learn more when the present at TFD but for now here’s what I can gather around DriveScale’s solution.

Basically they take a group of physical servers and desegregate these into pools of both compute and storage resources.  Converting these resources into what they call “Software Defined Physical Nodes” allows DriveScale to use both software and hardware to present these resources to our applications, with the ability to grow and shrink these resources as needed.  When the time comes to scale out we aren’t faced with the same challenges of purchasing pre-defined nodes where compute and storage come married together – Instead, we can leverage DriveScale to simply add more compute by bringing more physical nodes into the pool, or add more storage by importing in a bunch of commodity JBODs.  In the end, we can scale up or down as much compute and storage as we need, without having to worry about things like data locality – because we have DriveScale sitting between our compute and storage resources.

drivescale-architecture

This is all made possible by a couple of hardware and software components – First we have the DriveScale Management Server and the DriveScale software agents – these provide the core functionality of the product, by pooling all of the compute and storage resources within the rack into logical nodes.  All data runs through a hardware appliance called the DriveScale adapter, which basically is a central connection point for all of our physical servers, through a 10GBe network.

There is certainly a lot more that DriveScale’s solution provides, things like High Availability and visibility – but before I completely make a fool of myself explaining on how this all works I’ll just leave this post off right here.  Here’s hoping I can learn a bit more how DriveScale technically does all of this at Tech Field Day 12- and hopefully convey that information back 🙂  That said, if you want to learn more about DriveScale for yourself their Tech Field Day 12 presentation will be live-streamed.  I’ll hopefully have it setup on my page here, or if you wish, keep your eyes on the official Tech Field Day 12 page.