Home > Data Center News > HP puts fault-tolerant NonStop on a blade
Data Center News:
EMAIL THIS

HP puts fault-tolerant NonStop on a blade

By Bridget Botelho, News Writer
16 Jun 2008 | SearchDataCenter.com

IT infrastructure news
Digg This!    StumbleUpon Toolbar StumbleUpon    Bookmark with Delicious Del.icio.us    Add to Google

Palo Alto, Calif.-based Hewlett-Packard Co. has developed a new fault-tolerant blade server, the HP Integrity NonStop NB50000c BladeSystem, that analysts say is good enough to rival an IBM mainframe. The new blade combines the high-availability features of HP Integrity NonStop servers with the cost and energy efficiencies of the HP c-class BladeSystem.

For more on fault tolerance:
HP Nonstop server targets smaller businesses  

Virtualization to popularize fault-tolerant servers?  

NEC to roll out new x86 fault-tolerant servers

The NonStop NB50000c combines standard, off-the-shelf c-Class blade enclosures and blades and adds fault-tolerant features: The blade server interconnect in the back of the chassis is based on the high-speed, low-latency interconnect ServerNet for high availability, while NonStop software enables a fault-tolerant environment; if a problem arises on one blade, the workloads on it are immediately moved to another to prevent downtime.

The HP offerings are designed for IT shops in industries like financial services that strive for zero downtime. "Fault tolerance is for applications that need to be up 24 hours a day and for people who need to deliver answers in real time," said Randy Meyer, the director of NonStop product management strategy at HP. "It is really N+1 failover."

"[HP Integrity NonStop BladeSystem] is the one architecture in the industry that really has a chance at standing up to, or even surpassing, mainframe reliability.
Jonathan Eunice
principal IT adviser, Illuminata Inc.

The servers also feature standard HP management tools, such as HP Integrity Integrated Lights-Out (iLO) 2 , HP Systems Insight Manager and the Onboard Administrator for HP BladeSystem, and are based on dual-core Intel Corp. Itanium chips (and are compatible with quad-core).Up to eight NonStop blades fit in one chassis, Meyer said.

In comparison, HP's rackmount Integrity Nonstop servers are 8 U high and house four single-core processors.

"The [Nonstop] operating software can scale across multiple processors however it can get them, so moving to a blade was a logical change," Meyer said. "The system is designed in parallel, so we can have up to 4,080 blades in a single-system image. No customers have gone that big yet, but we do have customers doing several thousand."

Jonathan Eunice, a principal IT adviser at Nashua, N.H.-based Illuminata Inc., said HP's decision to use blades for fault-tolerant servers is strategic and beneficial for users. "HP chose to use its BladeCenter design as the foundation, based on its strategy that all server designs should become more modular over time. The blade foundation gives it some nice attributes, like incremental add-ons or component maintenance, higher volume and, thus, better economics."

A credible mainframe replacement?
HP has touted the Integrity NonStop BladeSystem as an ideal replacement platform for legacy mainframes because it runs the same kinds of applications at similar or better performance and at lower cost, Meyer said, citing the Standish Group International Inc.'s recent "Trends in IT Value" report.

Submitted by HP to SearchDataCenter.com, the report indicated that the Integrity NonStop BladeSystem performs the same tasks as an IBM mainframe at 35% lower total cost of ownership and with up to a 50% decrease in cost per transaction.

Numbers in reports like this can be tricky, and most of the time, vendor claims that their servers are "mainframe class" aren't credible, but two analysts said that in this case the claims are legitimate.

"[HP Integrity NonStop BladeSystem] is the one architecture in the industry that really has a chance at standing up to, or even surpassing, mainframe reliability," said Eunice. "HP's fault-tolerant architecture has been around since the 1980s, and it runs large stock markets, financial transaction systems, telecom analytics and fraud detection systems, and other 'must always, always, always stay up' computing."

Ken Cayton, IDC's research manager for enterprise platforms, said fault-tolerant systems like HP's Integrity NonStop servers are being adopted in environments where mainframes once would have been the only answer.

"HP's fault tolerant systems are frequently used in mainframe environments or where typical mainframes fail to meet availability and reliability requirements," said Cayton. "The new blade systems will have the same capabilities in a new form factor that should be more cost-effective and modular than previous systems while also being more energy efficient."

On the other hand, Eunice said, "HP NonStop and IBM System z use extremely different system designs and styles of operation, and as you might imagine, have different strengths, application libraries, and best workloads. They aren't 1:1 competitors like HP ProLiant and Dell PowerEdge; you can't run software for one on the other, say. But they do target many of the same kinds of applications: large transaction processing hubs, say, or large data analysis applications."

Still, HP has put its money where its mouth is; the company is so confident that the new Integrity NonStop NB50000c BladeSystem is as good – and better – than IBM mainframes that HP has offered the system plus a full year of NonStop platform software to customers who choose to migrate from mainframe architectures. Offered through the IT service company Logica, the incentive program, is called NonStop FREEdom.

Fault-tolerant systems on the rise?
Cayton said that data centers' broader use of virtualization for consolidation and backup makes high availability more vital and may increase the use of fault-tolerant systems.

But it's not a virtual environment that drives the need for fault-tolerant servers but rather the need for constant uptime. "Increasing transaction volumes [from things like] social-scale use of the Internet … and increasing use of data analytics for fraud detection, business decision making, risk analysis, etc. The underlying use cases are rapidly growing."

Other companies that offer fault-tolerant servers include Stratus Technologies and NEC.

Availability and pricing
The HP Integrity NonStop NB50000c BladeSystem is available today. Pricing starts around $300,000 for a chassis, two blades, drives and memory, and an I/O and power package. A fully configured system runs from $750,000 to $1 million.

Let us know what you think about the story; email Bridget Botelho, News Writer.Also, check out our news blog at serverspecs.blogs.techtarget.com.



Tags: Blade server trends and adviceHigh availability in the data centerVIEW ALL TAGS

Digg This!    StumbleUpon Toolbar StumbleUpon    Bookmark with Delicious Del.icio.us    Add to Google


RELATED CONTENT
Blade server trends and advice
Data center managers plan for power density jumps
Cisco's hopes its Extend Memory technology will boost UCS
HP's new server aims for cloud, Web 2.0 apps
Cisco Systems adds rackmount server option to UCS
Fujitsu launches new blade system: News in brief
IBM/Brocade deal reeks of Cisco hostility
Lowering data center cooling costs with airflow modeling and perforated raised-floor tiles
HP's Bladesystem Matrix to challenge Cisco's Unified Computing System
Cisco finally offers Unified Computing System details
Sun unveils new servers; former Sun exec launches startup

High availability in the data center
Apple updates Xserve, Symantec ships DR monitoring software: News in brief
Gmail outage caused by data center rolling blackout
Is Uptime Institute's data center tier system worth it?
VMware vLockstep update
Top disaster recovery budget wasters
Uptime Institute expands data center tier rating system
Alcatel-Lucent whittles costs with data center consolidation
Disaster recovery shortcomings addressed by Continuity Software
Colocation company 365 Main loses power

Mainframe computers
The mainframe is 45 years old
Using CICS dump tables to manage problems in online systems
High mainframe software costs may lead to platform's demise
Manage CICS workloads with transaction classes
Run CICS in batch to beat a shrinking batch window
Analyst group disses Hewlett-Packard report about mainframe migration
Using External Call Interface (EXCI) to access CICS
Mainframers go for a jog at Share user group conference
Mainframe student anticipates a bright future
A review of z/OS 1.9 features

RELATED GLOSSARY TERMS
Terms from Whatis.com − the technology online dictionary
blade server  (SearchDataCenter.com)
brick server  (SearchDataCenter.com)
Cisco Catalyst Blade Switch 3020  (SearchDataCenter.com)
compaction  (SearchDataCenter.com)
cook-off test  (SearchDataCenter.com)
server blade  (SearchDataCenter.com)

RELATED RESOURCES
2020software.com, trial software downloads for accounting software, ERP software, CRM software and business software systems
Search Bitpipe.com for the latest white papers and business webcasts
Whatis.com, the online computer dictionary



Efficient Management for Data Centers
HomeNewsTopicsITKnowledge ExchangeTipsBlogsMultimediaWhite PapersEvents
About Us  |  Contact Us  |  For Advertisers  |  For Business Partners  |  Site Index  |  RSS
SEARCH 
TechTarget provides technology professionals with the information they need to perform their jobs - from developing strategy, to making cost-effective purchase decisions and managing their organizations' technology projects - with its network of technology-specific websites, events and online magazines.

TechTarget Corporate Web Site  |  Media Kits  |  Site Map




All Rights Reserved, Copyright 2005 - 2009, TechTarget | Read our Privacy Policy
  TechTarget - The IT Media ROI Experts