Will You Be ‘Smarter’ in 2012?

The following PowerUp blog entry was written by Nigel Fortlage, vice president of IT, GHY International.

GHY International was recently named a smarter computing reference client. As I write this, I am thinking about what this means for my company.. Is it just marketing hype? Is there any substance to it? Why would you even care? Do you know what it means to be a smarter computing client?

By now I would think everyone knows about IBM’s Smarter Planet marketing initiative--a combination of systems, software and services, including analytics software, that allow smarter decisions to be made. Behind this concept are real-world solutions implemented by customers. Unfortunately GHY isn’t at this stage yet, so we can only speak to being a smarter computing reference client.

The smarter computing role, though, is less well known at this time. It simply means as IBM says, “The IT Infrastructure that enables a Smarter Planet.” You must have a smarter infrastructure that enables you to build a Smarter Planet solution. So let me share with you some insights to what it means to have simply smarter infrastructure.

If you click here, you can see and hear GHY's Executive VP share his insights along with mine as to why the solution is good for our organization. It’s a good way to understand from a business perspective what it means to us as a company using “Simply Smarter Infrastructure.”

I would like to summarize the key points for you that we outline in that video, as it is because of these points that IBM sees us a “Smarter Computing” reference client;

1. Save time--The solution we started implementing in 2003 has evolved through the years. Now you can implement an IBM Power server and get all of the benefits that we have right out of the box. If you’re fighting with disparate systems that don’t communicate well, have fluctuating bouts of reliability, or have created solutions so complex that you use multiple systems to deliver a single solution, then you too can benefit as we did from an IBM Power server. This changed our behaviors by saving 95 percent of our time that was spent servicing the data center rather than our users down to 5 percent.
2. Create solutions--What do you do with 90 percent of your time back? Focus on the business, I hope. We did and created solutions to solve problems and improve the working lives of associates in our organization. Imagine what it might mean for to find a single solution that saves everyone 10 minutes a day. With more than 80 associates, this resulted in 2.5 full-time equivalents (FTEs) saved for each solution we find.  We have found many dozens of solutions that help our organization today.
3. Virtualization choice--Imagine a single platform that allows a choice of more than 15,000 applications (AIX, i, and Linux) that leverage investment in the infrastructure but also don’t limit the business.

While it isn’t an inherent part of this story on smarter computing, I must stop here briefly to explain that we are also using VMware for Intel technology-based virtualization (Windows and Linux). My team keeps telling me that it’s different from the IBM Power Systems way. They prefer the IBM Power server’s single-level storage. To achieve anything similar to the Power results, VMware adds an extra layer of virtualization management via the VMware Control Center. From our experience this layer adds no value just complexity in managing the environment.

Our choice to use IBM’s Power Systems hardware allows us to leverage that infrastructure to maintain existing costs (leasing, licensing), reduce ongoing maintenance expense (hardware and software) and add new capabilities such as running x86 Linux workloads in an IBM Power virtualized server. That makes for simply smarter infrastructure.

What smart choices will you make in 2012?

Involvement in Local User Groups is Worth the Effort

The following PowerUp blog entry was written by Bruce Guetzkow, who has been an IT professional for more than 25 years, mostly with Power Systems servers running IBM i and its predecessors. His published articles cover utility programs and commands he created over the years. Bruce is currently Webmaster for the Wisconsin Midrange Computer Professional Association (WMCPA), an advisor for the Programmer/Analyst Track at Gateway Technical College in Kenosha, Wis., and an independent consultant for GmanTech Consulting in Delavan, Wis.

I've been a volunteer with the Wisconsin Midrange Computer Professional Association (WMCPA)  for several years. On March 20-22, 2012, WMCPA will be holding its 27th-annual Spring Technical Conference at the beautiful Grand Geneva Resort and Spa in Lake Geneva, Wis. The location makes for a reasonable drive for people from all over Wisconsin and Illinois and as far away as Michigan, Indiana, Iowa, Missouri and Minnesota.

If you've ever done any event planning you know many aspects must be coordinated: the venue, food, speakers, topics, lodging, audio-visual needs, entertainment and registration, to name a few. No one person could adequately deal with all of that.

Fortunately WMCPA has a five-person Board of Directors, elected annually, and four Conference Committee members (including myself) appointed by the Board that share in the work. Sharing the load makes these manageable tasks. I have learned a great deal over time about the people on our Board and Committee. Everyone has unique talents that they bring to the table to enhance the process. Whether it be effectively communicating with potential speakers, negotiating with the different venues and sponsors, creating banners, composing documents or just putting a few links on a web page, everyone contributes in his/her own fashion. We also get help from IT students at conference time thanks to our special relationship with Gateway Technical College in Kenosha, Wis., and their IBM i-based Programmer/Analyst program.

Recently a few of us got together to work on “the grid," which is the document that lays out specific days, times, speakers and topics for our two-day conference based on topics submitted by potential speakers. As of this writing it is about 95-percent complete. There will certainly be a few changes, rearranging times, adding or removing particular topics as needed, but there should be enough information for IT professionals to make up their minds as to which sessions are of utmost importance to them. If it isn't available on our website, it will be shortly.

For the last few years we have also been offering a bonus day (Tuesday) for anyone that registers for the two normal conference days (Wednesday and Thursday). That's when we host the Tuesday night roundtable discussion where several of the conference speakers discuss a topic of interest. The Wednesday Casino night entertainment allows attendees to play games with fellow attendees and speakers in a relaxed atmosphere.

When first approached to volunteer for WMCPA I must admit that I was a bit reluctant. After all, did I really have time to do all of that work? Well, it's true that there is work to be done, but I quickly realized that I am only one of many, and many hands make light work. I have also developeed friendships over the years with the Board and Committee members, and many of the WMCPA members-at-large that make me want to volunteer year after year.

Plenty of free webinars and other online events are available, but when you want to really learn about a topic there is nothing quite like a technical conference where you can talk directly with industry experts. If you're planning on being anywhere in the Midwest we hope you'll consider joining us at our 27th-annual Spring Technical Conference. If not, check out a local user group in your area for their monthly or conference offerings. And, while you're at it, consider volunteering--it's time well spent.

Flattening the Learning Curve for HPC with Linux on POWER

The following PowerUp blog entry was written by Chandler Wilkerson, one of Rice University’s Linux cluster administrators in the Research Computing Support Group, where he administers the IBM POWER7 Blue BioU cluster and the university's campus Linux deployment infrastructure.

When last I wrote, I detailed the user education challenges in optimizing the running of HPC codes on our small POWER7 cluster, Blue BioU. Quite a bit has changed from a year ago. This May, we completed an expansion of the Blue BioU cluster by 30 nodes, increasing the compute node count from 18 to 48. At the same time, we upgraded the operating system from Red Hat Enterprise Linux (RHEL) version 5 to version 6, which gave us huge gains in support for the POWER7 hardware.

The most notable feature  has been a recently incorporated set of patches to the upstream Linux kernel (and incorporated into RHEL6 kernels) that added asymmetric SMT packing for multi-threaded Power CPUs. This has greatly simplified our handling of basic codes by improving default performance characteristics of Linux applications on POWER7 hardware.

Before this patch, application processes would be randomly distributed across threads, sometimes putting multiple execution threads on a single core, while leaving other cores empty. This would lead to throttling of the overloaded cores, not to mention thread contention, all of which combined to provide horrible performance in HPC codes where synchronization ensures that the whole system cannot run any faster than the slowest process.

The asymmetric scheduler now knows to migrate CPU intensive tasks to the fastest threads on individual cores and to spread out the load over all the cores. We still recommend using CPU pinning techniques to achieve optimal performance, but now the performance boost is closer to 5 to 10 percent instead of multiple orders of magnitude.

RHEL6 makes another important change in that the entire OS is now compiled in 64-bit by default, as opposed to a 64-bit kernel with a mostly 32-bit OS stack. Perhaps more importantly, GCC now outputs 64-bit code by default, so we don’t get nearly as many library architecture conflicts now as we had when everything defaulted to 32-bit.

Before the upgrade, even moderately experienced users would need special attention to get their codes running reasonably on Blue BioU. Post upgrade, compiling and running jobs on the cluster is much simpler. The proof is in higher, more efficient utilization statistics and reduced problem ticket load for us admins.

Insights from the Road: A Week with Power Systems and Linux

The following blog entry was written by Nigel Fortlage, vice president of IT for GHY International. He is the architect and overseer of GHY’s technology backbone and infrastructure, oversees relationships with the company’s technology-related vendors, and is visionary in equipping GHY to meet its present and future challenges.

The temperature wasn’t the only thing that was hot when I visited Austin, Texas, and LinuxCon 2011 North America in Vancouver, Canada, this past week. I took to the road to share insights from GHY International, which uses the powerful combination of Linux, the open-source operating system, on IBM Power Systems in a production environment. I also wanted to get a sense of the current state of Linux on Power hardware.

From a corporate perspective, GHY has invested in state-of-the-art technology that allows it to seamlessly communicate with its customers and customs offices in the United States and Canada. It also prides itself on providing its customers with the human touch.

“We offer the best of both worlds: As companies look at their options for trade services, they’re looking for partners that can deliver all the important technology-based solutions but also maintain close personal relationships and accountability for performance,” says Reynold Martens, executive vice-president.

GHY International is in its second decade of using Linux in the enterprise, the last nine years on IBM Power hardware. Approximately two-thirds of our server environment consists of virtualized machines running on our IBM Power 750 server. We use SUSE Enterprise 11 Linux on our server instances.

We run very diverse workloads from core network services including a firewall on an IBM Power LPAR, to email and Web services, and finally to an application-development platform based upon the industry-standard LAMP (Linux, Apache, MySQL, PHP) stack.

Using this powerful combination we have created solutions that save the organization hundreds of thousands of dollars annually. These savings are calculated in real hard dollar savings from licensing, maintenance costs, training, etc., to FTE savings. As an example managing email, spam, anti-virus, context filters, etc., saves us an estimated 9.5 FTE annually.

It’s because of our use of Linux that I attended LinuxCon an event of The Linux Foundation, which is a non-profit consortium dedicated to the growth of Linux. This year’s North American event was held in Vancouver, BC, Canada, from Aug. 17 through Aug. 19.

Of special note during the LinuxCon North America 2011 event was a keynote with Dan Frye, vice president, Open Systems Development, IBM Systems and Technology Group; Jon “Maddog” Hall, executive director, Linux International; and Eben Moglen, director-counsel, Software Freedom Law Center. The keynote was moderated by Jim Zemlin, executive director, The Linux Foundation. From a commercial position it was the feeling of the group that Linux is where it is today because IBM stepped up and made its original $1 billion commitment back in 2000.

News of the Week was poorly timed when Phil Robb, director of the Open Source Program Office at Hewlett-Packard presented his keynote the morning of Aug. 18, just as the news broke from HP corporate about the company’s withdrawal of WebOS based products, The audience, which was plugged in, was aware of this news before the keynote even wrapped up. Unfortunately for HP the keynote was all about how the Linux technology-based WebOS was so great. Needless to say, HP’s booth at the event was bare after that.

I did some recon at the event to understand how the market sees the IBM Power platform’s position. That the “Jeopardy!” playing computer, Watson, is an IBM POWER7 processor-based system running Linux that speaks volumes to the masses, IBM has a home run with Watson.

Currently, Red Hat and SUSE are the IBM enterprise partners that provide distributions for Power Systems. Red Hat is the North American market leader in x86 Linux workloads, while SUSE, the new/old standalone business unit from the recent acquisition of Novell by Attachmate leads in non-North American markets as the choice for enterprise Linux. During a conversation, a SUSE executive indicated that the company's figures indicate that it also accounts for at least 50 percent of the IBM Power market and more than 80 percent of the IBM System z marketplace. Of special note for me is the fact that historically SUSE has been in step with IBM Power hardware releases.

For those wondering, I did have very detailed discussions with SUSE executives about the whole Attachmate deal. While a number of issues that reassure me of SUSE’s future in this market were discussed, such as Attachmate is investing, not divesting of this product portfolio, some nagging questions still need to be answered as to the long-term future impact to the open-source marketplace and potential indemnification that vendors will still need to put in place for enterprise users. If anything I think it plays very well for SUSE, which holds perpetual rights to use all products and their intellectual property without license.

Surprises and Delight

It was with great surprise and delight that I found some interesting tidbits that the IBM Power community should be aware of.

Gentoo, which is generally regarded as the Linux distribution that is like getting a box of Legos without the instructions for those who like to build things, has a community build of its distribution for IBM Power Systems.

Canonical, the distributors of Ubuntu, also have a community build for IBM Power Systems available. Since Canonical is already in the business of selling enterprise distributions, maybe the company will consider a future opportunity to make its community build an enterprise offering.

The last tidbit is more a future opportunity than a solution for today, maybe. IBM’s love/hate relationship with Oracle is no secret. It’s also widely known that IBM Power Systems with enterprise Linux can be a good solution for Oracle application users. Oracle, of course, bought MySQL, the “M” in the LAMP stack abbreviation. IBM has offerings to migrate to DB2 (a great database) but that breaks the industry standard of the LAMP stack, and at this time I am not aware if the DB2 team has done anything to make DB2 more of a drop in replacement to MySQL. The original developers of MySQL now have an offering called Maria DB, only available in an x86 distribution directly, but it is a direct binary replacement to MySQL, thereby preserving the LAMP stack and requiring no re-development while migrating to a different database, away from future concerns of what Oracle may do with MySQL. I found it interesting that this may not be as hard as it sounds as Gentoo’s public distribution already contains the IBM Power version of Maria DB. This is definitely one case that requires some R & D to understand what it can do and again increase choice in the IBM Power market.