The Itanium Solutions Blog

For the electricity industry in India, managing customers is all about power and availability.

Yes, handling the ever-increasing energy production and distribution demands in a rapidly growing economy is a challenge, indeed. Along with that, however, comes a sometimes less discussed but no less critical problem: how to handle the explosive growth in IT infrastructure to manage the growing volume and complexity of its customer base.

In Maharashtra, the country’s third largest state in area and home of its largest city (Mumbai), this problem is massive. The good news is that business is booming, with over 15% of the India’s industrial output produced in this region and the state boasting India’s highest per capital income as well.

Power consumption itself is putting major strains on the region in many ways, as demand currently exceeds supply on a regular basis, and management of customer issues is also growing at a staggering rate. But thanks to Itanium and HP, Maharashtra Electricity State Distribution Company Ltd. (MESDCL) has found a way to handle both current IT demands more effectively and to establish a system with built-in capacity for expansion.

Every month MESDCL handles 15 million customers and the expected growth rate is at least 500,000 new customers to be added every year. Its annual revenues are currently just under $5 billion (U.S.), and the company itself has over 75,000 employees. To support its substantial client base MESDCL had originally written its billing system in Cobol, a workable system for a time but with significant limitations on managing large data arrays.

As capacity demands began to build significantly in the 1990s, MESDC realized it needed to replace its older software solution with a more modern one based on Oracle, which it completed in 2000. Although that helped, as customer growth continued MESDC quickly realized the hardware part of the solution also badly needed upgrading.

As just one example, MESDCL’s existing PA-RISC system, even with the power of Oracle behind it, took as much as 24 hours just to process 10,000 bills. There were other issues as well, including the need to handle a unique photo-based approach to recording electrical meter readings, something that added even more data management problems and was completely unmanageable within the existing IT environment.

MESDCL needed a system with much faster overall response times, lower IT costs that would scale even more efficiently as customer demand grew, and dramatically improved productivity on a grand scale.

Their final solution, which involved a complete re-engineering of their IT infrastructure (including a new picture-processing capability for the photo-based meter readings), used 12 HP Integrity rx3600 servers and 10 HP Integrity rx6600 servers, all driven by dual-core Itanium microprocessors.

With this solution in place, MESDCL has seen its billing processing time for those 10,000 bills drop from 24 hours to only one hour. In addition, because of the increased processing speed, MESDCL was able to roll out its new photo-based meter reading systems for approximately half of its 15 million customers – in the first six months of system deployment.

The solution also offers a far more energy-efficient hardware environment than before, along with a less costly IT support environment as well. Both will provide MESDCL with the “power” to scale their mission-critical computing needs for many years to come.

With the economic problems around the world right now, you might wonder how that is affecting the IT industry — and the server market in particular. Would companies retrench and hold back on spending? The answer is that, with results counted from the first six months of an admittedly challenging 2008, the server business is apparently doing quite well. And Itanium fared well, too — with results that may suggest some of what is happening behind the scenes.

The results, which were summarized in a recent survey released by Gartner, Inc., include an interesting statistic: that worldwide server shipments in Q2 2008 were up 12.2 percent over the same quarter last year. Revenues were up as well, but by a smaller amount, 5.7%. This more-than-reasonable performance in a tough economic climate was driven by a combination of a “continued upswing in x86 server replacements” plus “Web center build outs” and growth in emerging markets.

That refers to the entire market, but how did Itanium fare in the same period? From one respect, raw shipments, not as well, with Gartner reporting that “RISC-Itanium Unix Servers” saw a 7.9 percent drop from the same quarter a year earlier. However, revenues for this server category were growing at a rate of 9.4%. As Jeffrey Hewitt, research vice-president of Gartner said, this indicated “that higher-end systems were the hardware platforms that drove sales in this space.”

One conclusion behind all this is that, even in a tough economy, IT executives continue to make the decision to upgrade their systems to increase productivity through performance improvements, higher availability and lower cost-of-ownership. Good tactics, under the circumstances.

The investment in higher-end Itanium systems at a considerably higher rate than in previous years suggests a more strategic move is also taking place as well. With major enhancements introduced for Itanium this year with emphasis on Virtualization, plus a number of developments that make Itanium-based server solutions more dynamically configurable on the fly, my guess is that you are also seeing an increased rate of server consolidation for companies that can afford to do so. All of which positions those companies even better when the overall economy begins its upswing, hopefully sometime in the near future.

Not a bad strategy, and the revenues are high enough to suggest that far more than just a few good companies are making this change. It is something for all IT organizations out there to consider as the economy continues to navigate the current doldrums and charts a course for a stronger recovery path ahead.

So stay tuned. My guess is that data for the third quarter will show even more of this trend.

The Itanium Solutions Alliance recently announced it reached agreement with Sophos (http://www.sophos.com/), a leading provider of enterprise solutions for IT security and control, to provide a new level of support for Itanium systems running on Linux.

Sophos already supports HP-UX, OpenVMS and Windows (including the recently-released Windows Server 2008) on Itanium-based systems, but with a new release of Sophos Anti-Virus coming in Q4 2008, it will now support Itanium-based systems running Red Hat EL 4.x and Red Hat EL 5.x versions of Linux.

Along with the continuing growth of Linux for mission-critical enterprise applications is, unfortunately, a parallel growth in the attacks on such systems with advanced computer viruses and malware. So the porting of Sophos Anti-Virus to support the latest Red Hat Enterprise Linux editions for Itanium is welcome news indeed.

Sophos’ approach to its Enterprise Security products is highly-regarded worldwide and with good reason.  One of the keys to this is a unique approach to what they refer to as “Behavioral Genotype Protection”® and run-time protection algorithms. These systems track not only the presence of known viruses and malware but also their artifacts as well — in registry changes and other data log alterations, all with the end result of identifying the presence of such dangers as early as possible.

Like other companies providing antivirus software, Sophos delivers proven algorithms to identify, quickly disable, and quarantine antivirus threats. But where they are different is their extensive focus on enterprise systems. They understand the importance of providing powerful control tools for setting up and managing the overall IT end-to-end security infrastructure, including centralized installation services, malware alerting mechanisms, and advanced security software appliances to monitor web gateways to your system in real-time.

The combination of these tools is a sophisticated approach to isolating both known and new malware and system threats at their earliest emergence. The tools are accurate, thorough, and run quickly with minimal IT overhead.

Add to this that the new offerings Sophos will provide are some of the most powerful and extensive anti-virus solutions for Itanium-based systems running on Linux, and you have a compelling product any of you Itanium Linux Users out there should be considering.

Watch for it in Q4 2008.

It is now prime vacation time for many and travel planning is at its peak. When you make your plans, it’s very possible you’ll be making your reservations via an Itanium-based solution.

Whether you book online, by phone, or in person, there are generally two parts to the system: an online booking engine and a company which provides the connection between that booking system and the travel providers. That second part of the system is called a Global Distribution System.

The largest global processor of these travel transactions is Amadeus Global Travel Distribution S.A., which supports more than 94,000 travel agency locations and 32,000+ airline sales offices in hundreds of countries and territories. To do so, they must manage 500 million travel bookings per year, through connections to 95% of all scheduled airlines, 22 car rental companies serving 36,000 locations, nearly 77,000 hotels, and 17 cruise lines.

On a daily basis, Amadeus handles nearly 300 million transactions of all kinds and up to 2 million internet-based bookings at their Erding, Germany data center, one of the largest civilian data centers in the world. With price competition becoming incredibly fierce, Amadeus’ Fare Quote system is in high demand. This allows customers to compare prices based on different carriers, routes, and dates before booking. How do they make this happen? At the core of their system is a growing server farm with 59 HP Integrity rx8620 servers, each of which are equipped with 16 Intel Itanium processors, running SUSE Linux and Amadeus’ Altéa reservation system. Thanks to extensive Amadeus-exclusive IT infrastructure innovations, SUSE, and the Itanium-based systems, they are able to handle more than 5,500 consumer requests per second during peak hours. The system actually takes less than one third of a second to process data inquiries and completes each fare search within five seconds.

All of this processing is an incredibly complex task. For example, when a customer is searching for the lowest airfare possible, the Amadeus solution can produce more than 200 low-fare options per query. These options have to be accurate, timely, and structured so customers can easily navigate options based on changing dates, switching airlines and airports, and even reconsidering destinations. This means providing many millions of instructions to the server system per query.
These demands make system response time, rapid reallocation and scalability critical elements to the server system.

Fortunately, for Amadeus, and probably you if you’re booking your next trip online, all of these computing needs are met with Itanium-based systems.

This collaboration continues to bear fruit in many ways.

At its simplest level, the companies provide the many building blocks of hardware platforms (too many to mention), OS choices (from Linux to customized Unix to Windows), and mission-critical enterprise applications for financial and customer management (and others).

Another way is through careful coordination by multiple partners to optimize a solution to best benefit a customer. Great examples of this include:

• The Scottish Life Division of Royal London Group migrated more than ten million lines of code and nearly a billion data records to a new Itanium-based system, thanks to the leadership of MSS International, a system migration specialist, MicroFocus Studio and Server for legacy applications issues, HP for its HP Integrity, HP 9000, StorageWorks Enterprise Virtual Array Systems, and HP OpenView, and Oracle with its 9i Database, Application Server, Oracle Management Pack, and Oracle Data Guard. The end result improved batch job run times by one-third and cut the total cost of ownership by 50%.

• Kindred Healthcare responding to its ten-fold increase of database size over eight years (to two Terabytes) by moving to an Itanium-based platform resulting from the combined efforts of HP with its Integrity rx7620 and rx4640 servers, SAP with its mySAP ERP 2005 solution, Microsoft Windows Server 2003 and SQL Server 2000, and other custom enhancements. Online user response time was improved from 1.3 to 8.0 times (depending on the application and number of users); report generation improved by more than two times; and batch processing was 24% faster.

A third way is through the rapid development and deployment of specific new applications which leverage the latest advancements by other Alliance members. Some of the recent enhancements involve Intel Virtualization Technology for the Itanium architecture at the hardware level itself and are now being fully supported by Red Hat Enterprise Linux 5.1 and NEC, Fujitsu, Hitachi, and HP, among others.

Finally, a conscious team-oriented collaboration helps produce innovative and breakthrough products. One of the best examples is the recently announced Itanium-based solution featuring dynamic partitioning, full hot-adding of processors, memory, and I/O host bridges for Windows Server 2008.

Ultimately the partnership between Itanium Solutions Alliance members and you is probably the most important. It’s those relationships between you and your providers that are the links between what you want to achieve with your computing and how you’ll achieve those goals. Dr. Carlos Simmerling of Stony Brook University, the winner of the Alliance’s Humanitarian award in 2007, discussed that a key reason why his biomedical research modeling performed so well was due to hands-on application and integration support from Alliance Founder SGI. SGI’s team offered valuable insight which helped his application run as quickly as possible. Thanks to the collaboration between SGI and Dr. Simmerling’s team’s efforts in the modeling area, we may be significantly closer to a cure for HIV.

Sometimes we need to take a step back to recognize how important teamwork is in today’s computing world. Indeed, how vital are the individual players in your computing environments vs. a collaborative approach? Do you trust your IT decisions to someone outside your department or are you picking and choosing the players for your own “Dream Team”? What do you expect to gain from working with a community of hardware and software vendors?