Article Rating: |
||
| November 11, 2008 10:00 PM EST | Reads: |
3,898 |
EMC unveiled its massively scalable cloud storage Monday, the stuff that it’s been working on for a long time now under the code names Hulk and Maui, Hulk for the hardware and Maui for the software, branding the whole megillah Atmos as it moved out the door. Some features like compression and de-duplication that are usually add-ons are thrown in for nothing.
EMC unveiled its massively scalable cloud storage Monday, the stuff that it's been working on for a long time now under the code names Hulk and Maui, Hulk for the hardware and Maui for the software, branding the whole megillah Atmos as it moved out the door.
The stuff's for sale but EMC won't say how much it costs or share any performance specs. It says the price is competitive. It's silly, EMC suggested, to post a list price when, God knows, nobody pays list. Hmmm.
Some features like compression and de-duplication that are usually add-ons are thrown in for nothing.
Being cloud storage the widgetry's early adopters have reportedly been petabytes types - either they've already got that much data on hand or soon will have. EMC didn't say who they were but suggested they're Web 2.0 providers, telcos and service providers and media and entertainment houses. At least that's who Atmos is supposed to appeal to.
EMC says Atmos is different from anything it's sold before or from anything anybody else is selling now, a claim that puts HP's nose, for one, out of joint.
HP figures its previously announced NAS-style ExDS9100, its so-called "Extreme" storage, which it now says will be out before the end of the month, is the stuff of clouds. HP Extreme runs $2 a gigabyte, which sounds reasonable enough although a fully tricked-out system costs $1.5 million.
EMC claims its stuff is a different species of beast than SANs, NAS or CAS, the latter being content-addressable storage in case you haven't been keeping up.
It's supposed to be a new storage category, an example of cloud-optimized storage (COS) largely because it's policy-based, not to mention Internet-sized and catering to unstructured data, the stuff most people are socking away these days.
Atmos is special enough that EMC's got special sales teams targeting key opportunities and specialized deployment teams backing them up. The company is promising a broader rollout next year and continued enhancements.
Atmos is currently available in three configurations: 120TB, 240TB and 360TB and EMC imagines users scattering any number of these racks - the bigger two outsized 44Us - around the world. No matter how many there are they're still just one Atmos system - (think unified namespace) - and there's no theoretical limit to how many multi-petabytes you can string together at how many hundreds of sites.
EMC says 200TB-300TB can be added at a time without a flap complements of the namespace.
The policies see to it that the right information goes to the right place at the right time. And Atmos' reliability, accessibility and performance trick is in how many copies of the data there are.
This is not RAID sort of stuff. It's object-based and the objects (documents, videos, digital images and music) are tagged with metadata. Atmos can handle billions of these objects, which is another way of saying it's policy-based.
EMC says not to expect the usual kind of management. That stuff's old hat. Those conventional concepts just don't exist in Atmos, which was designed to be administered by fewer people than one would expect.
It's an auto-config, auto-healing architecture that allows multi-tenacy.
The hardware is standard x86 Intel gear; the software's Linux-based but proprietary, which gives HP something to scratch at. Its stuff is supposed to be more open.
Published November 11, 2008 Reads 3,898
Copyright © 2008 SYS-CON Media, Inc. — All Rights Reserved.
Syndicated stories and blog feeds, all rights reserved by the author.
- Is the PR Business Extinct? Yes
- Government IT & Cloud Computing: Themes for Discussion
- GovIT Expo Highlights Cloud Computing
- Cloud Computing Best Practices
- VIP Invitation For the GovIT Panel October 6, Washington DC
- Forget Defining Cloud Computing
- Why SOA Needs Cloud Computing - Part 1
- The Cloud Transition: What Does It Mean For You?
- IBM Puts Systems Chief on Leave of Absence
- Cloud Expo and the End of Tech Recession
- Oracle Fined for Sun Ad
- IBM Exec Out on Bail as Galleon Sinks Below the Waves
- Is the PR Business Extinct? Yes
- The Difference Between Web Hosting and Cloud Computing
- Government IT & Cloud Computing: Themes for Discussion
- GovIT Expo Highlights Cloud Computing
- Cloud Computing Best Practices
- The End of IT 1.0 As We Know It Has Begun
- VIP Invitation For the GovIT Panel October 6, Washington DC
- The Case for Single-Purpose Services
- Product Evaluation: JBoss TCO Calculator
- Forget Defining Cloud Computing
- Why SOA Needs Cloud Computing - Part 1
- Cloud Expo Power Panel on SYS-CON.TV
- Java vs C++ "Shootout" Revisited
- Where Are RIA Technologies Headed in 2008?
- WebSphere Application Server Java Dumps
- Breaking News: New Internal IBM Report Says "Another Flawed Study"
- Last Exclusive JDJ Interview With "IBM's" John A. Swainson, Now CA's Newly Appointed CEO
- How To Deploy Scalable WebSphere Applications Using "Maven" Build Tool
- Your Guide to Portal Clustering in WebSphere Portal Server 5.1
- Developing Java and Web Services Applications on Rational Application Developer V6
- The Top 250 Players in the Cloud Computing Ecosystem
- Automated Deployment of Enterprise Application Updates
- Putting IBM's WAS On Unix - WebSphere Application Server
- Profiles for WebSphere Application Server 6.0
































