IBM Lobs Benchmark at Sun
IBM has fired a shot across Sun's bows in an attempt to win big in the retail sector
December 22, 2004
![Network Computing logo Network Computing logo](https://eu-images.contentstack.com/v3/assets/bltde8121fc52c5c8f3/blt3f3d0318f746b1c2/65a530e4187606040a1d8b8c/placeholder.png?width=1280&auto=webp&quality=95&format=jpg&disable=upscale)
There is little cheer between IBM Corp. (NYSE: IBM) and its Unix archrival Sun Microsystems Inc. (Nasdaq: SUNW) this holiday season.
IBM, which is never one to miss a marketing opportunity, has fired a shot across Suns bows. The hardware giant has teamed up with supply chain specialist Manugistics Inc. to test the vendor’s NetWORKS Fulfillment application on its new pSeries 590 Unix server.
The method in IBM’s madness? Sun tested the same application on its own Sun Fire 6800 Unix machines last year. And here's the surprise: IBM says it has blown the doors off its rival’s performance figures (see IBM Outperforms Sun).
The benchmark volley comes at an opportune time. For both IBM and Sun, selling hardware into retailers is big business -- and this is the time of year when servers come under the most strain.
In the test, a single 590 running 32 Power5 processors was capable of handling 38.4 million Stock Keeping Units per hour, according to IBM and Manugistics. In plain English, this means that the machine can "ring up" nearly 40 million retail items per hour, like an enormous cash register.Last year Sun ran a NetWORKS Fulfillment test on two Sun Fire 6800 servers running a total of 48 UltraSparc processors. The machines managed around 30.5 million SKUs per hour.
But Gordon Haff, senior analyst at Illuminata Inc., is somewhat skeptical about this type of thing. “You should always take benchmarks with many pinches of salt,” he warns.
Haff believes that there is much more involved in buying a major piece of kit, like a server, than a single benchmark. “Firms will base their purchase on a series [of benchmarks] and even on a custom benchmark based on their specific [situation],” he says.
Nonetheless, he acknowledges that, in this case, IBM has got the edge on its rival. “I don’t think that it’s much of a surprise that IBM, with its Power5 processor, performs a lot better than the Sun UltraSparc-based servers. Power5 is significantly faster than UltraSparc.”
Even though Sun used an earlier version of the NetWORKS Fulfillment application, the results could also prove significant for customers outside the retail sector. With many software vendors, particularly in the database space, charging on a per-processor basis, improving performance while using fewer chips could be a major boon.Sun did not respond to NDCF’s requests for comment for this article, although IBM was more forthcoming on its Power processor plans. A spokesman told NDCF that a Power5+ processor is already in the pipeline: “There will be performance improvements next year,” he says. A Power6 processor is expected sometime in the 2006 timeframe, he adds.
IBM’s Power5 message appears to be getting through. Today, the hardware giant announced that Lockheed Martin Corp. has consolidated a number of Unix servers onto its pSeries machines (see Lockheed Martin Picks IBM P5 Systems).
— James Rogers, Site Editor, Next-Gen Data Center Forum
You May Also Like
Maximizing cloud potential: Building and operating an effective Cloud Center of Excellence (CCoE)
September 10, 2024Radical Automation of ITSM
September 19, 2024Unleash the power of the browser to secure any device in minutes
September 24, 2024