Why RISC it? Scale-up Industry Standard x86 Servers are here TODAY!

[#1]

hosted for Colin England

HP Worldwide Marketing Campaign Manager

We’ve all heard about it. Many of us are already doing it. But for those who haven’t yet made the switch, shifting applications from aging RISC/UNIX-based platforms to x86 systems is probably looking like a more and more attractive IT strategy for your business. And for good reason.

Why? For so many applications and workloads running on RISC systems from 2008 or earlier, the evolution of x86 in recent years has made it harder to justify keeping those apps in their UNIX environments. New, scalable, industry standard servers, paired with an ever-hardening open-source OS, like RHEL, can now deliver the performance and reliability those apps demanded when  you deployed them on RISC 3 years ago, but for a whole lot cheaper. Of course, this assumes the business end-usage of those workloads has remained generally the same.

The truth is, you made a good call choosing POWER/AIX or SPARC/Solaris to run your important stuff a few years ago. x86 just wasn’t there yet. Actually, it wasn’t nearly the platform you could rely on, for a few reasons:

• It was not resilient enough
• Performance was lacking
• It had limited scalability

But this is no longer the case. Let’s look at some benchmarks to prove this out.

On the TPC-H benchmark (tpc.org), RISC/UNIX systems from the ’06-’08 time frame racked up around 100K queries per hour. A newer, scale-up x86 system hit closer to 200K QpH on the same test.

If we look at this same benchmark and focus on price-performance, x86 really begins to shine.

Current benchmarks show x86 system at $2.68 per query, vs. today’s RISC system at $20.60. That’s almost 1/8th the price-per-query. When we compare to the RISC systems of a few years back, it’s 1/20th. Scalability is growing too. With Intel’s new Westmere processors, that same 8-socket HP ProLiant system can now scale to 80 cores, and 4 TB of server memory – making Moore’s law look more like a standard to exceed rather than just an objective to meet.

And of course, it’s not only about performance and price-performance, but…. can you really trust industry standard platforms with your vital apps and databases? More and more so, yes you can. New developments like self-healing resiliency, made possible by a resilient system fabric, maximizes uptime for workloads on, for example, a new  8-socket HP ProLiant server, adding 200% more availability than HP’s previous 8-socket box.

To be sure, leave the truly mission-critical workloads alone. They still need the stability of RISC. But for those marginal apps, and maybe some databases, it’s likely becoming just too expensive to keep them on a RISC/UNIX system, especially if you’re looking at an expensive upgrade or license renewal. And with the services offerings from many IT vendors and SI’s becoming more and more adept at these sorts of migrations, your business risk is dropping.

So if you’re not already considering a shift to open standards for some of your important workloads, it might be worth a closer look.

Download the Whitepaper: “Price. Performance. Advantage. Red Hat Enterprise Linux on the HP ProLiant DL980 Server.” And watch the free webcast.

hosted for Colin England

HP Worldwide Marketing Campaign Manager