The secret behind the unique Power servers: a look under the hood

comarch
Comarch Data Center, Krakow, Poland. Source: Comarch

[adv] IBM Power servers have guaranteed performance and reliability for many decades, but some persistent biases persist. What is so unique about IBM Power Systems and how can your organization deploy them? We consulted IT specialist Comarch.

The first IBM Power server dates back to 1988 with the legendary AS400. Since then, the product line has undergone several technological and name changes, appearing under the current name Power Systems. The promise has never changed all these years: the servers should stand for performance, robustness and reliability. Polish IT company Comarch therefore offers its own cloud services on stable Power infrastructure.

The largest banks, insurance companies and retailers have IBM Power servers in their IT arsenal, and for good reason. Yet there are also (unjustified) preconceptions about the role of IBM Power Systems in modern IT environments. What makes IBM Power Systems unique, what can your organization do with them, and why does it pay to hire an integrator to help with the migration? We explain it all to you in this article together with IT specialist Comarch.

From AS/400 to Power Systems

We’ll start with a brief history. The ancestor of all IBM Power servers is the AS400, which saw the light of day in 1988. The AS400 simplified installation and maintenance by combining compute, storage and I/O into a single server. The server was much more compact than the mainframe boxes common at the time.

What was even more unique was that each file or device was considered an object. The user-friendly software (OS400) and high reliability and security were also strong strengths of the AS400. To this day, high reliability, high availability and maximum security are central.

In the 2000s, servers underwent several refreshes. The OS400 operating system was replaced by IBM i. In 2008, IBM merged the System i and System p product lines into Power Systems, the current name that sticks to the servers. Comarch built its own PowerCloud over that to make everything even more performant.

RISC

Under the hood is the Power microprocessor. The tenth generation was launched in 2021 and is baked on a 7nm process. IBM puts up to 15 SMT8 cores and 30 SMT4 cores in a socket. Per core sits up to 48 (instruction) and 32 (data) KB L1 cache and 2 MB L2 cache and eight execution segments that can process up to 20 instructions.

The sum is in the parts. Each Power10 processor has two halves, each consisting of eight SMT8 cores and 64 MB of L3 cache. One core always remains disabled, effectively giving you fifteen operational cores and 120 MB of L3 cache.

The underlying Power ISA architecture also deserves a word of explanation. The foundation is laid with RISC(Reduced Instruction Set Computer). RISC differs from CISC in that it is much less complex: each instruction performs one function per clock cycle. The common x86, also in the server industry, is CISC at heart; another well-known RISC implementation is the ARM architecture. RISC is gaining importance today because it enables more efficient and economical hardware.

No legacy

Nevertheless, IBM Power Systems must contend with some persistent prejudices. Because the oldest servers have been around for several decades, they are sometimes labeled legacy. Legacy has acquired a negative connotation in the IT industry today and is crudely equated to “old junk you want to get rid of.

This is certainly not true of IBM Power Systems. The servers have been able to stay relevant for so long just because the hardware and software have evolved with the times. Why would the largest banks risk their daily operations by using obsolete technology, knowing that any downtime would cost them tons of money?

Don’t worry in that regard: IBM Power Systems servers are equipped to meet any of today’s IT challenges, no matter how complex.

Power Cloud: it doesn’t have to be expensive

Another prejudice about IBM Power Systems is that it is an expensive technology. In addition to purchasing the servers, they also require maintenance and specific knowledge and expertise in the IT team. If you don’t take proper care of servers, performance degrades over time. Indeed, it can be costly if you take this maintenance into your own hands as a user. Power System experts are also rather scarce in the job market.

Partnering with a reliable partner can help reduce Power Systems costs. Polish IT company Comarch offers cloud services on its own IBM infrastructure. Comarch’s experts take server management out of your hands, so the IT team can focus on what your organization’s real needs are.

Comarch’s PowerCloud also provides flexibility. No one wants to have too little computing power available, but overprovisioning is at least as damaging. With Comarch, you make sure you always have the right amount of compute and storage capacity, and you can scale up when needed.

Finally, Comarch PowerCloud is synonymous with security. In the data center, both the server infrastructure and the network are secured with robust infrastructure and the technical expertise of dedicated experts. So you don’t have to worry about inefficient or inadequately secured IT infrastructure: you get the quality you pay for.


This is a commercial contribution in partnership with Comarch. For more information on the PowerCloud-solutions.

newsletter

Subscribe to ITdaily for free!

  • This field is for validation purposes and should be left unchanged.