What is our primary use case?
I use IBM Power Systems for online transactional processing, and in parts, also for online analytical processing.
We mainly use IBM Power Systems' IBMI, which is an operating system with an integrated database. For that system, the database is just another file system, which is ingenious because it saves you a full database administrator, or multiple database administrators, and that's a fairly cool thing.
What is most valuable?
They are reliable, their availability is very good, and their scalability is second to none. I can go from really small to very, very large, and still stay on the same platform, still go with a single server, rather than using multiple boxes and clustering them. To give you an idea, when considering Lotus Notes, that was an email workload, workflow software from IBM. When I ran that on a Windows X86 server, I could do something around 1,000 users per server. Beyond that, I needed additional servers. A couple years back, we introduced the same system for the US Department of Defense, and we had 275,000 users running on one and the same hardware, running on one machine.
There are multiple layers to the reliability aspect of IBM Power Systems. The hardware itself is very reliable due to the fact that I only have a very reduced set of hardware, so driver problems are pretty rare. I need to support two or three different ethernet drivers here and not thousands of ethernet drivers as with Intel. The hardware is very redundant, so high availability or disaster recovery becomes a breeze because normally, I just end up calculating one room, one local system, one remote system.
There are several features of IBM Power Systems that significantly improved processing efficiency. These machines come with built-in server virtualization, which has been there for the last 20 years. We started in 2005 with that. This allows me to put multiple workloads on the same machine, and those multiple workloads can communicate with each other very fast. In a normal network, I often encounter latencies because I need to travel through many cables, and I have switches in between that may not be available right away for forwarding my signal. With an IBMI, or with a Power System in general, I have a virtual network, meaning from one partition, I copy my data into the RAM, across to the next partition, and then put it through a network driver again. This is a latency-free network.
I am just starting with AI workloads and initiatives that are supported by IBM Power Systems. The reason for that is that IBM was slightly slow with supporting this. For commercial workloads, I wouldn't see anyone doing that before the end of this year because OpenShift Red Hat OpenShift for AI will not be available much earlier than that. I've been playing with that in a lab environment for about a year now, and it runs quite well.
What needs improvement?
There are definitely areas for improvement, but that's not really from the technological side of things. I rarely find any people with Power skills on the market. When I get young people from university these days, all they can do is Intel, which is everywhere. They can't do any DB2; all they can do is Oracle because that's what they've learned in university. I'd IBM to do more in the university space to attract more people with specific skills.
From the features perspective, we're really looking forward to AI co-processors such as the Spyro cards announced for later this year. We would be very interested in those having the capability to also train models on the platform, which would really make it a great AI platform.
What was my experience with deployment of the solution?
If I know what I'm doing, it's very simple to set up IBM Power Systems. But I've seen people struggle; it's somewhat complex as I've got to think in very different levels. For example, I need to set up my hardware management console first, then set up my service processor because the HMC is the Dynamic Host Configuration Server for that. If I set up the machine first and the HMC second, the addresses will go to the wrong address range. But those are typical beginner's mistakes that can be avoided by reading the manual. IBM's documentation is great on that, so it is easy yet straightforward.
What do I think about the stability of the solution?
Their stability, reliability, and scalability are the main benefits that IBM Power Systems bring to the table, improving the way my organization functions. They scale from almost desktop size up to major mainframe levels. They are very stable; even on a single system, I can easily get 99.999% uptime. Unscheduled downtimes are very rare. They have features such as alternate processor retry, where the processor computes something, then notices due to a self-test that this can't be right. When I need eight bits, it provides 10 bits, incorporating one detection bit and a correction bit. If I notice one bit is off, out of those eight, I use the correction bits to replace that one bit. Even if RAM goes haywire over time, the machine doesn't hang; it first replaces that with a replacement bit. If that's not possible anymore, the machine will automatically disable that particular RAM region, so I have slightly less RAM, but the machine will still run stable.
What do I think about the scalability of the solution?
They are reliable, their availability is very good, and their scalability is second to none. I can go from really small to very, very large, and still stay on the same platform, still go with a single server, rather than using multiple boxes and clustering them.
How are customer service and support?
My experience with IBM technical support really varies. Their level one support consists of personnel collecting data. They can get tiresome by asking redundant questions, which are probably going to be redundant for the supporter I'm going to end up with, but they ask them because they're in the list. With the second level support, it really depends on where I end up. Usually, if I end up in the US, the support is very good. However, if I end up in Eastern Europe, which is another typical support region because it's cheap, the support gets worse than in the US. In East Asia, the support gets abysmal. I try to get my support calls pre-routed to the US; I'd rather wait a couple of hours for support becoming available there than going to support somewhere in the Philippines or Indonesia.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
I've worked extensively with Lenovo systems and Dell systems, and I've always preferred IBM Power Systems.
How was the initial setup?
The main pros of IBM compared to Lenovo and other solutions start with the mechanical engineering of those systems, which are usually well thought out. There are many hot plug capable components, and components are easy to reach. It's easy to get somewhere; usually, I can slide them out of a rack without disconnecting the cables if they were mounted the correct way. IBM spends a lot of effort on stuff other manufacturers don't prioritize. The big con with IBM is the price. If I look at a project purely from a total cost of acquisition perspective, IBM is usually very expensive. But when I consider total cost of ownership, or how much that thing is going to cost me over the first three or five years, then IBM is really far ahead of the others. However, the moment I buy that system, it's not all that apparent that I'm saving on two database administrators.
What about the implementation team?
IBM Power Systems require hardware maintenance and software maintenance. But if I manage that, they're actually quite easy to maintain.
What was our ROI?
I don't have metrics to share, but I have seen excellent ROI from IBM Power Systems.
On average, for smaller systems, it usually takes about six to twelve months to see the benefits from IBM Power Systems. For larger systems, it can be either very fast because I can do things on there that I just cannot do on X86. But if I'm looking just for the cost benefits, that can take anywhere from a couple of months to a couple of years.
What's my experience with pricing, setup cost, and licensing?
The big con with IBM is the price. If I look at a project purely from a total cost of acquisition perspective, IBM is usually very expensive. But when I consider total cost of ownership, or how much that thing is going to cost me over the first three or five years, then IBM is really far ahead of the others.
Which other solutions did I evaluate?
I would advise other organizations considering IBM Power Systems to think before they buy, plan before they buy, and calculate before they buy. When I buy the right machine for what I'm trying to do, it's a great machine. But I shouldn't compare apples to oranges; for instance, if I have a 20-core Intel system running somewhere, to get the equivalent processing performance, I do not need a 20-core IBM Power System. The equivalent would be more a five-core IBM Power System or a Power System with five cores activated. So, don't compare apples to oranges. If I say a 20-core IBM Power System costs 100 grand, and a 20-core Intel server costs me 50 grand, this is expensive, it's not because it's actually four times the performance per core compared to Intel. So I need to be smart.
What other advice do I have?
Our experiences so far are that IBM Power Systems is a very good platform for running AI workloads. It's not yet a great platform for training large language models, but then I train once in a great while and run them all the time. So for having a decent runtime environment or a decent governance environment for my AI, it's brilliant.
It's very good at integration because it sticks to the norms, sticking to the RFCs. There are sometimes problems with other vendors in the market such as Microsoft because they do not stick to the norms, or even Cisco because they stick to the final draft usually to be on the market fast. But even with that, I can work around it, so it integrates quite well.
I rate IBM Power Systems 10 out of 10.
Which deployment model are you using for this solution?
On-premises
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?
Other