You’ve Still Got Friends In High Places
September 23, 2019 Timothy Prickett Morgan
For the past 25 years, and across a total of 78 meetings over that time, the Large User Group, or LUG as it is known, has been helping Big Blue keep the AS/400 and IBM i platform on the right track for the very large enterprises that have the system as their platform for running the mission critical applications that are at the heart of the business. The LUG has put forward thousands of requirements to IBM – both things that the organization wants IBM to do and other things that it wants IBM to stop doing – and a pretty significant percentage of these are implemented.
While the LUG is not for everyone, there are some spaces available at the moment and we wanted to remind you that provided you have the right profile, you too might be a friend in high places in the IBM i ecosystem and use your influence to help drive the platform forward into the future.
The LUG started somewhat informally back in 1994 at the COMMON user group when IT executives at six big AS/400 shops got together and all realized that they had big boxes and that they also had a totally different class of problems from most of the AS/400 midrange. So they formed a group and hosted meetings with each other to talk shop. In 1995, they came up with a list of 11 handwritten requirements for the AS/400 platform that they handed Big Blue. This was the at the height of the transition from CISC processors to PowerPC AS RISC processors and also from 48-bit to 64-bit OS/400 – a time of great change in the platform and burgeoning sales. In 1999, the LUG incorporated as a business in Minnesota and started charging an admission fee to be in the group. Today, the 100 members of the LUG pay $3,000 a year to be in the organization and they have three meetings a year at IBM’s Rochester Labs. The fees cover the costs of administering the group and paying for the food and beverages at the meetings; members pay for their own travel and lodging when they are attending.
The LUG can handle 115 members and currently has 100 companies subscribed, so there is a chance that you might benefit from being part of the organization. In the past, membership was gated by the amount of raw capacity, in aggregate CPWs, that the company had deployed. A little more than a decade ago, companies had to have a total of 30,000 CPWs to be able to join and a few years later that was bumped up to be 50,000. Considering that a Power9 core is delivering somewhere between 18,200 CPWs and 15,500 CPWs depending on the configuration of the system, that is no longer a lot of CPWs in the scheme of things. And to give you some perspective, the current 100 members of the LUG have an aggregate of 53 million CPWs of performance across their IBM i iron, which works out to an average of a half million CPWs each or a factor of 10X more than that entry bar to apply for LUG membership.
To get a more representative metric to screen for admission, and to try to bring in service providers who don’t have big machines as much as they have lots of small ones, the LUG switched to counting LPARs. To be precise, you have to have at least 30 IBM i licenses spread across no more than three enterprise-class Power Systems machines to at least 60 IBM i licenses spread across at least a dozen Power Systems entry machines to apply. The latter criterion is clearly aimed at service providers. By the way, membership is contingent on participating – you can’t just sign up and not go to the meetings or not provide feedback to IBM through the LUG. You can’t just have passive influence – you have to employ active influence. So don’t sign up unless you can commit to the three meetings a year and spend the time away from the shop.
To get a sense of what is going on in the LUG these days, we had a chat with Dutch and Dave, two members of the board of directors for the LUG whom we highly suspect of not using their real names. Which is fine because the whole point of the LUG is to be able to speak freely amongst each other and to IBM without having political or personal blowback. Our thanks to Lynne Benedict, the business analyst in charge of the group, for setting this up.
Timothy Prickett Morgan: It’s been a while since I talked to anyone from the LUG, which is probably not a good thing. So my apologies on that front. My first question is how important is the LUG to the high end of the IBM i market? Do you have more or less influence than you had a decade or two ago, and how do you wield it?
Dutch: I will give you a good example from a few years ago. IBM had decided to get rid of the HMC and make it available as a virtual machine on a Linux LPAR on IBM i. And we convinced them that they didn’t want to do that because that would be onerous within our companies, and pulled the plan from doing that.
Here’s what IBM didn’t appreciate at the time. Adding a Linux LPAR to systems in a large shop causes lots of other work to be done – by people outside of our organization, in some cases. There’s mandated software that has to be on a Linux LPAR, there’s configuration mandates that may or may not conflict with what they were looking to put in for the virtual HMC and then there is the fact that none of us had any expertise with how this thing was going to be supported. What do you think, Dave?
Dave: I would say that, overall, being in the LUG is still very beneficial. We still have a good view of the roadmap, so we see what’s coming. We also have the ability to influence product directions in the future. With some of the stuff in the latest 7.4 release, we had several folks involved in a team that was reviewing it all ahead of time and throughout the whole development process. All that has been very beneficial. But to me the key benefit is when IBM delivers technology, we are able to adopt it and use it because it’s what we asked for and what we had influence on throughout the development cycle.
TPM: I don’t expect you guys to tell me anything that’s on the roadmap, but I am curious how far out you can actually see with clear visibility what IBM is planning. I think five years in the enterprise computer business is about as far as anybody can think about or see with any sort of accuracy. But do you have a sense of what’s happening with the processors, the memory, the storage, the interconnects – plus all of the stuff in and around the operating system releases – many years out? How far out do they let you see?
Dave: They let us see out as far as seven or eight years in some cases. However, as it gets closer you get to see some more detail. Those very early views are very high level, strategic goals and they are really asking us if this or that is something that would interest us. And then as it gets to that five year window, then IBM asks us for some real input, and then at three years we actually get some pretty good visibility on what a product is going to look like. It has been very beneficial to the group as a whole to have such a long view, to help prepare a business for stuff when it does get delivered.
Dutch: I think that it’s me that I think DB2 Mirror is a great example. You know we’ve been talking about that for years. We have been seeing demos for at least the prior 18 months before it was launched and we have had discussions talking about the user interface for it. I have been making decisions about how we’re going to use the product – and efforts we are not going to undertake – based upon the fact that we know this is coming along. That’s probably one of the most valuable things about the LUG to a large organization: We’re not going to spend time with an alternative to DB2 Mirror that we would have had to basically throw away because we know it is coming. At our shop, we are using a combination of Power HA for disaster recovery and software replication from one of the independent vendors and we will be eliminating the software replication and adding DB2 Mirror.
TPM: Talk to me about the demographics of the LUG members. Do these companies tend to be from North America, or North America plus Europe?
Dave: We have around 25 international members at this time and about 75 members from the United States. This can obviously change over time, but we cap the membership at 115 members companies today. And that’s due to the logistics of the meetings. If we had a massive influx of 20 customers that wanted to join we would probably review that and see what we could do logistics wise to accommodate. But today it’s capped at 115.
TPM: Do member companies tend to be fairly current on hardware and software? I assume there are not members that are lagging back on Power6 or Power7 hardware and on anything older than IBM i 7.1. But maybe that’s not the case.
Dave: I wouldn’t say that. We range from Power6 up through Power9. There are some people that lag behind due to trying to schedule upgrades, which is time consuming in a big corporate environment. So there’s still a few stragglers on Power6. From an operating system perspective, there’s still a few stragglers back on i/OS 5.4 – and those are very few – all the way up to 7.3 and we will soon have people on IBM i 7.4.
Dutch: It’s interesting. Because member organizations are so large, we do run the gamut, but it is in the single digits when you are talking about the number of members who have LPARs that are on 5.4 and 7.1.
What drives us for the most part as the LUG is the fact that we need to maintain that supportability for our organizations because we are growing and dynamic organizations. If you’re if you’re a small mom and mom and pop shop that you’ve got one guy running the system and he is supporting a small pizza box system, it may not matter to you what’s happening in the future. You don’t change often enough that you need to predict the future and more importantly you have a much smaller chance of hitting something unexpected in your IBM i platform as you support the company. So these smaller companies can stay on an unsupported release if they have to. In my company, I wouldn’t have my job if we went down and the reason was because was we were on an unsupported operating release from IBM. So, in the LUG typically, unless it’s on an LPAR that is not mission critical, you don’t see anybody on an unsupported release.
Dave: It typically starts with the application. To go and replace an ERP system sometimes takes years and years of planning. I think as a whole, LUG members have come pretty far forward. I think the number that are on releases older than IBM i 7.1 and the number that are on machinery older than Power8 hardware are few and far between. It’s a handful, not a lot.
Dutch: And typically, even those will tell you they’ve got a project scheduled in an upcoming quarter to move ahead. Or there is some reason they can’t. It could be they have a piece of critical software and the vendor has gone out of business, or the application vendor doesn’t support the current release.
TPM: What kind of cadence are you guys comfortable with for the hardware upgrades and the software upgrades. It seems to be we’re on a three-ish to four-ish year hardware cadence and it’s looking like we are on a three-ish cadence for IBM i releases. Steve Will, the IBM i system architect said recently that four years is too much distance between releases and two years is not enough. Are you happy with the way this stuff rolls out? Is it consumable to you given the amount of change you’ve got to manage whenever you introduce a new system to the to the datacenter? And then is it good or bad to have these things in a sort of lockstep?
Dutch: The Technology Refreshes have made a big difference. At my company, we are at the point where we will put IBM i 7.4 on after it’s been out for a little while. With everything else we are still at IBM i 7.2. We made the conscious decision to skip IBM i 7.3. Part of the reason for that decision was the fact that it takes a lot of time and effort to roll out a new release. And not just from our teams, but also the application team. And the other thing was that, at least for us, IBM i 7.3 didn’t have a lot of new stuff to justify the upgrade. We could get a lot of functionality in IBM i 7.2 by using the Technology Refreshes, whereas it didn’t used to be the case.
As for hardware upgrades, our guiding factor on how often we replace the servers is how long the finance department is going to depreciate the machines. Each company balances that. We don’t replace the server until we fully depreciate it, unless there’s something that drives us to that. When we add new servers, we always add what’s latest on the market. In some cases, we’ve delayed a new purchase because we knew a new release and new hardware was coming out. The last thing we want to do is buy a piece of hardware in the month before the next one is announced.
Dave: I would concur. I think the three-year to five-year cycle seems to fit pretty well what LUG members are doing. At my company, we try to depreciate our hardware in three years.
TPM: If you could change anything that IBM is doing, what would it be?
Dutch: I wish IBM would have done a better job of marketing the platform. That’s probably the biggest battle we face. Someone fairly high up in the company asks me: “Why are we using that old hardware?” And then we have to go through an education process of bringing them up to speed on why it is not old hardware.
I will give you one of the best examples of this. I was forwarded an email from several echelons up in the company and the comment was: “I’m concerned that we’re running the business on an AS/400.” I replied to the person that said this: “I’d be concerned if I was running the business on an AS/400, too. Fortunately, we’re running on an IBM i on the latest Power on the latest IBM Power hardware. So we don’t have to worry about that.” But it’s that perception that IBM years ago – for whatever reason – decided it wasn’t marketing the platform. And now, most of the growth in the platform that you see is overseas.
One last thing: We have a membership application out there, and the admission is not strictly about size. If you are good sized but maybe don’t meet the requirements from the LPAR standpoint but you’re doing open source stuff, you’re using RDi for everything, if you are moving forward on the platform, those carry just as much weight – well, almost as much weight – as how large you are.