Sometimes the best advice you can find is from watching what your peers are doing.
Back in March, HelpSystems released its first annual IBM i Marketplace Survey report. 344 of our peers in the IBM i and Power Systems world explained how they use the platform, highlighting the good, the bad, the challenges, and the concerns. This is great news, especially as an annual survey, because it's great to know what other people are doing with Power Systems and IBM i and how they're doing it. You can download it free from the MC White Paper Center.
I'm going to try to look at the results from a slightly different angle, highlighting topics of the report in bold. While I think the study overall puts the state of IBM i in a good light, I'm a real data hound and I want to see much more! Hopefully, next year's report will give us an even better look and dig a bit deeper.
There's no real surprise in the The Data Center: POWER Server Capacity section. I'd argue some of this study is working through semantics. For the chart, "How many IBM Power servers do you have?," I would imagine they mean how many Power Systems servers. A POWER6 System i is still a "POWER" server, but it's not a Power Systems server. There's a difference between POWER and Power, where the latter implies Power Systems and that's if the respondents designate between the two. If that's the case then only 1.5% of the respondents are running hardware older than 2008. But that's not the case based on the Staying Current section of the report, where 9.6% have some POWER4, 34% have some POWER5, and 25% have some POWER6. I think the reason there's a discrepancy is because customers aren't designating between POWER and Power. It's important to do so. All Power Systems is POWER, yet not all POWER is Power Systems. POWER can mean an actual iSeries, which of course is a different brand and a different hardware architecture altogether.
38.4% of respondents indicated that they have a single Power server running their business. The designation between POWER/Power is negligible only when you think about it as an IBM midrange server and the value of having one machine run your business. Almost 40% is a good number.
In terms of partitions, 62.7% of respondents have multiple IBM i partitions. I'm very interested in this number, but I'd want to take it a step further to know "how many production IBM i partitions do you run" to see how the numbers shift a little bit. How much would that change the 35.5% who run 2-5 partitions?
Furthermore, I'd love to know how many of these shops run a Hardware Management Console (HMC). If partitioning is on the rise, then what tools are people using for partition management? I got my first HMC this year, and it's probably the most valuable component of my POWER8 upgrade. It opens up so many doors for partitioning and systems management.
With regards to AIX and Linux on Power Servers, only 6-7% run AIX and/or Linux on the same hardware as IBM i. My question arises in the 30.2% who run Linux on other platforms. The study doesn't state whether that's Linux on Power Systems or Linux on x86-64 servers. I'd assume it's a mix of both but would like that broken down next year to see the dispersion. We do know that 37% of respondents use Linux alongside IBM i, so I'd suggest that Linux on Power is far more used than people think.
Outsourcing is always a fun topic for me. What's neat is that 74.4% of respondents do not plan on looking at outsourcing to a Managed Service Provider (MSP). It's a testament to the fact that IBM i is very easy to manage.
Yet outsourcing IBM i is not a bad thing. If you have an IBM i partition in the cloud and it works for your business, then more power to you. MSPs can offer you things that you might not have the resources to do in-house. I would argue that even I, a staunch promoter of in-house IBM i, use MSPs by way of contractors. My Power Systems servers are in-house and we manage it for the most part, but I regularly have consultants working on it in areas where we're not too strong or because we lack the time.
Regarding Staffing, there's no real surprise on these slides. It's more proof that IBM i is relatively self-managing. Just over 80% have two or less administrators. Almost 70% have two or fewer data center operators, with 33.7% having none! On the Automation slide, over 50% have some sort of notification tool in place rather than staff manually checking dashboards or looking at disk status or the system operator queue.
In terms of developers, it's encouraging to know that about 30% have 3-5 developers in-house. The Modernization slide gives a good breakdown of what languages are used for new development. RPG being the biggest at 86.6%, with Java (37.8%), .NET (20.9%), and PHP (20.1%) getting fair share. In terms of interfaces, we're talking semantics again, with the question about if the main application is green-screen. 9% say no and that their primary interface is web-based. Almost 8% say no and that their primary interface is graphical. Well, web-based is graphical. What I'd like to see is how many are using graphical screen scrapers vs. building a modern GUI. The web vs. thick-client GUI stats would be interesting to know next time. Either way, the key to take out of this slide is that almost 40% do not have green-screen interfaces being used on their main application. That's fabulous! I personally didn't think it was that high.
In terms of Mobility, I'm a little surprised that only 28% are actively building applications for iOS and Android. Whether that's building applications designed to be responsive to mobile device web browsers or actual installable apps is a little unclear. What is clear is that 72% give a flat no to having a mobile computing initiative.
In the spirit of modernization, the Web Servers slide is highly positive. It shows that about 70% run web servers on IBM i. In reality, everyone runs a web server on IBM i unless they manually disable the IBM HTTP Server *ADMIN job from running. What I get from this slide is 70% do active, intentional web serving on IBM i. Personally, I have about 25 web servers on one partition and two web servers on my other. IBM i does web well. It's great to see that promoted here.
In terms of Return on Investment, almost 94% state that IBM i gives a better ROI than other operating systems. I'd love to have the names and numbers of the other 6% to understand their reasoning. It's amazing that, even at a 94% approval rate, I (and probably a good number of you) think 94% isn't high enough. It's an amazing ROI. If you're among the 6%, then shoot me an email in confidence because I'd honestly love to see your math.
Given the high amount of ROI, an interesting slide to see is on Platform Migration. I think those particular results need to be taken with a grain of salt, just based on wording. The question posed is "does your company plan to leave the IBM i platform?" Right away, I'm concerned that this question is a lightning rod for some of the naysayers who will go unnamed rather than give them any press. It's easy to take this question and put a negative spin on it. People can look at that slide and negatively promote that 45% are either talking about leaving or are leaving. That's absurd. The positive aspect of the slide is that 11% are planning on leaving within five years, or an 88% retention of customers who are giving the platform a 94% approval rating based on ROI.
I would've used positive language and said "Do you foresee IBM i running your business for x years?" with x being a time value and "migrating away" being one of the options. You'd probably see a result that's closer to the reality of the actual platform migration statistics in the survey results.
I reached out to Steve Will, Chief Architect – IBM i Operating System, for comment. I assumed the actual "people leaving the platform" number was actually even lower; somewhere between 2-5% which would not even account for people joining the platform community.
"As far as I know, we don't know how many customers leave the platform, and we have no way of knowing. People generally don't tell their vendor when they stop buying a product, and there's no mechanism, at least that I know of, for us to know it happened.
Your guess of 2-5% even seems high to me, given what I see. I often get called in when a company is considering changing platforms. I suspect when a business is doing this sort of evaluation, it would look to a lower-level employee like "we plan to leave IBM i in the next X years" because they hear their bosses talking about the possibility.
However, I have never had a customer actually leave once I get a chance to talk to their leaders about the direction and value of the platform. Businesses which look at the possibility are sometimes serious, or sometimes they are just doing "due diligence" to make sure the platform should remain in their long term plans.
Now, have I HEARD of customers who have left? Yes. I have heard of them. I know a couple, and I hear, anecdotally, of others. But the number seems small.
I have also heard, at least as often, of clients who put projects in place intending to move off of IBM i, but those projects fail and they remain on the platform, because they cannot adequately replace the stable, low TCO, scalable platform which IBM i provides them. These customers come back to IBM i, but they don't trumpet their failure to move away. Sometimes, the clients become great references for us, because they see how bad it would be elsewhere. But they don't talk about having tried to leave.
Finally, while the data in this question is useful, it does not cover businesses which have been added to our customer base."
Another challenge is that the person answering the survey may be aware of a technical component related to their job but may not be involved with the strategy at their company with regard to IBM i. Alison Butterill, IBM i Product Offering Manager, explains that IBM has "no idea who answered these questions. It might have been anyone in an organization—a systems administrator, a programmer, the CIO, a manager, a .NET administrator, the CFO, the VP of Operations. The person answering may not be aware of the IBM i value. That is shown by the 11% who 'don't know.' I would think that a company should re-examine their platform choices on a regular basis. They should—it's part of good business practices. Therefore, someone not understanding what is happening and was asked to do just such a thing may also be interpreting that request as a decision to leave IBM i on Power." This is a great point. Imagine if you were tasked to ensure due diligence with regard to your infrastructure options but not necessarily briefed on just why you were doing so. It may inadvertently put a doubt in your mind about the future of the platform at your company. This is a strategic question posed in the same set of questions about IASPs and programming languages. Not that a programmer wouldn't have the right information with regard to strategy, but it's a strategic question only a few slides away from discussing disk space. The potential for error is just as great as a CIO talking about the different web servers you use.
While most of the survey is truly very good, I think we need to be careful with some of the statistics pulled, most specifically the Platform Migration section. While customers do come and go, speculation of future events need to be avoided in favor of available facts.