Since 1986 - Covering the Fastest Computers in the World and the People Who Run Them

Language Flags
September 16, 2010

The Fate of Oracle’s Grid Engine: To HPC or Not to HPC?

Miha Ahronovitz

The growing concern in HPC circles about the fate of Oracle Grid Engine (OGE) after the merger with Sun in February 2010 is the catalyst for asking this rhetorical question.  

Since GridEngine (the open source project of Oracle Grid Engine) is part of Oracle IP, and Oracle has a valuable new asset that can facilitate the entry to new markets valued in billions of dollars. Yet OGE is now underutilized and overlooked.  Oracle strategy seems focused primarily  to monetize immediately new products from acquired companies. OGE monetary potential visible with the naked eye is misleading. The product itself and the know how of its’ engineers can create new solutions in Oracle impossible to acquire from other companies, because they don’t exist yet. One example is to help secure a pie in the emerging smart (power) grid market valued at $200 billions per year and exceeding the size of the Internet market.

Oracle has a treasure  in its backyard that must be extracted.  Oracle needs new eyes to look at what they already have and see what it was not apparent before.

For now the access to  Oracle Grid Engine product page  is complicated. If one does not know anything about what Oracle Grid Engine can do, it is highly improbable to discover this product by simply browsing Oracle’s web site.  The easiest, and apparently the only way, is to search for product  by name. New potential users can not discover Grid Engine by just browsing oracle.com site to  read the latest documents available: a new Oracle Grid Engine whitepaper, a High Scalability whitepaper and a a new data sheet.

OGE’s existence is not widely known, even within Oracle itself.  At a recent August 10, 2010 Next-Generation Datacenter summit entitled: Optimize with Oracle  event in Palo Alto, during the Oracle Cloud presentation, someone in the audience asked the same question I am posing here…”what about Grid Engine?” The speaker had no answer.

Douglas Eadline recently authored an article about Grid Engine in Linux magazine, in which he speculates:
 
“While GridEngine does not take any of Oracle’s products to task, the company may see wider application for it in the cloud space, which might mean more aggressive marketing—which of course might mean no more open source.

This brings us to the Shakespearean-like question Oracle will need to tackle: “To HPC, or not to HPC?”

At Sun there have been times when HPC was the most unpopular thing and times when HPC was labeled everywhere.  Most people are sure HPC (under whatever name or hype) will make a comeback – must come as a comeback – even at Oracle. Oracle must realize that Sun acquisition requires a new frame of mind, that enlarges the enterprise database frame of mind that brought so much success to Oracle.

HPC is where innovation happens in enterprises. Oracle can assist enterprises today in managing costs, employees, distribution channels, their sales force and all that good stuff. But creating innovative new products and solutions which in the long term form the basis of growth isn’t really connected to Oracle’s classical enterprise products.

A seasoned executive from HPC industry commented: “Add to it the convergence of the IT infrastructure operating enterprise and HPC services then you will see that Oracle blocks out an opportunity if they don’t listen to the HPC side of their customers and doesn’t give them suitable products.”

Can  Oracle  wait until they can buy some company which has shown how to bridge enterprise and HPC computing  rather than be the innovator themselves?  My answer is No.

Here is an  example of where Grid Engine can bring a competitive edge for Oracle. OGE can help Oracle being very competitive on smart power grid market. Perhaps they did not realize this yet. But they will.

The new $200B per year smart power grids has no dominating product ready. President Obama has called for the installation 40,000,000 smart meters and 3,000 miles of transmission lines. Venture Capital companies like  Nth Power,  Foundation Capital, New Enterprise Associates, Kleiner Perkins Caulfield and Byers, Siemens venture arm and many others are actively investing. Cisco just acquired a start up Arch Rock, a smart grid monitoring company. The market is there, but the products are incipient.
  
How important is the  smart grid market for Oracle ?  Larry Ellison himself is a featured speaker at Focus On Smart Grid section at Oracle World.

Smart Grid applications have elements of real time, as data are collected from millions and millions of wireless sensors requiring immediate processing and decision making. These streams of intensive data processing are familiar in HPC and Grid Engine – which since version 6.2 Update 5 has a Hadoop module to map reduce raw data before being placed in databases. The competition is tough, the pie is huge ($200B per year), more than the entire internet market. 

The future Oracle module needed to collect and pre-process real time the data in any power grid cloud application, will   in super-scale clusters, with enough resources to manage a national power grid. This is OGE turf’s.

There are no established players on this new market, and there is nothing to buy that works off the shelf. The competition is fierce and Oracle has  in-house something better than what’s out there to bring faster to market than the competition.

There  is optimism with Mike Hurd, ex HP CEO,  joining as co-President of Oracle. He has top exposure on hardware markets and ought to influence the Oracle culture in the right directions. The smart grids, are just an example of one of the fields where Oracle’s answer to the question from the title is “yes, let’s HPC”. 

Miha Ahronovitz
Principal Ahrono Associates