In The IBM i Trenches With: IBM Champion Ash Giddings
December 13, 2021 Timothy Prickett Morgan
It may be hard to believe, but not everybody who has worked on an AS/400, iSeries, System i, or IBM i platform is a programmer. Even if they care a great deal about programs and how they run on the boxes. Ash Giddings, who is an IBM Champion in 2021 and a newly appointed product manager at high availability software provider Maxava, is one such chap, and we got an opportunity to have a chat with Giddings about what is going on with performance management, systems management, disaster recovery, and high availability out there in the IBM i base.
Like many of you who read The Four Hundred (and those of us who write it, now that I think about it), Giddings has a career that spans decades, starting with a degree in data processing and business technology from Tresham College in Northamptonshire in the United Kingdom, north of London and halfway between Cambridge and Birmingham; he has been able to stay in this area for most of his career, which is a testament to the longevity and geographic diversity of the OS/400 and IBM i platform as well as to working remotely.
Giddings initially worked on the System/38 in the 1980s when it was a relatively new platform and an insurance company in England that was a mainframe shop had a System/38 running its fund management software and learned how to look after it. Like others at that time, he could easily land freelance gigs operating System/38s and then AS/400s after they came out in June 1988, and he did that for five years when “there was lots of money to be made,” as he put it. In 1993, Giddings became an operations shift manager for Bass Brewers – the one that makes that excellent ale – where he was responsible for managing four operators and had responsibility to keep fourteen AS/400s and one RS/600 at two sites that ran the brewery’s operations.
In 1996, Giddings moved to the European arm of timesharing operator RCI Group, which is part of Wyndham these days, as a senior midrange analyst responsible for two AS/400s, seven dozen Windows servers, and a few Unix boxes that were supporting operations in Europe. This is where Giddings did capacity planning and managed hardware and software upgrades as well as disaster recovery and replacing the lights-out management tools for these machines. Seven years later, Giddings joined Getronics, which set up an AS/400 i890 and offered managed services on top of it to customers as well as managing 20 other systems around the United Kingdom; and managed partitions, clustering using DataMirror iCluster, and provided Level 2 and Level 3 technical support for these platforms to customers.
In 2005, Giddings moved to Halcyon Software to become its technical services manager in charge of the three tech support teams in the United Kingdom, the United States, and Australia, and did a stint as its product marketing manager before and after HelpSystems acquired Halcyon in April 2015, and then took over the Halcyon product line for the conglomerate. Giddings stayed there until joining Power Systems cloud provider Skytap in June for a very short engagement, and in September joined Maxava as a product manager. Giddings has plenty of experience with HA solutions, and so moving to Maxava seems natural., and many former colleagues also work at Maxava, including Martin Norman, who is the manager of business services and operations; John Dominic, vice president of global sales; and Mark Scanlon, who heads up EMEA sales.
Timothy Prickett Morgan: Let’s talk about operations first because you started there. I have my own conceptions, and perhaps misconceptions, about how operations work at IBM i shops these days. Tell us how it really works. Are there a lot of programmer/ops hybrid employees? Are there dedicated ops people, whether they are in-house or a third party managed service or a consultant?
Ash Giddings: These days it is more and more common for people to not be dedicated to Power Systems. They have dozens of plates to spin on sticks, and IBM i is just one of those plates. Going back a maybe five or ten years, that wasn’t the case and there were dedicated people that had dedicated skills, but not there are definitely developers that are tasked with having to look after Power boxes. And, you know, developers being developers, they like to code things, they like to script things.
The IBM i platform becomes more and more integrated, probably due to the stability of the box and the longevity of the applications, and it sits nicely in the middle of it all and it is the crown jewels of many departments and everything talks to IBM i now. It has ended up being this big database server in a lot of companies, with everything else interfacing to it. At this point, it is only the very big shops that seem to have dedicated Power operations people.
TPM: That’s refreshing. Because if you look back 20 years, the AS/400 and IBM i was its own island with its own people.
Ash Giddings: I sit on the Common Europe Advisory Council, and we get to meet and greet IBM two or three times a year. And one of the common themes that that we keep talking about is making IBM i normal, so that it doesn’t stand out. Instead of letting it gather dust in the corner, make it integrated with everything else. A lot of the open source stuff added to the platform is definitely starting to make IBM i feel normal. The Navigator for i web interface, again, it feels normal. It’s quick and it’s game changing. Somebody can understand it and pick it up. It’s not this strange box is strange language that nobody else understands. That’s key for things going forward, to be sure.
TPM: Is the IBM i CL that is used for scripting much different from other scripting languages that are used, such as PowerShell on Windows or Bash on Linux?
Ash Giddings: On IBM i, I think CL is a little friendlier. I wouldn’t say CL is more forgiving, but there’s more help and more steering you in the right direction than maybe other scripting tools.
TPM: Let’s shift gears to performance. In the early AS/400 years, I made a good living obsessing about it, building all kinds of price/performance spreadsheets that showed every which way CPU, memory, and disk arm choices could alter performance – and at what cost. Every bit of capacity mattered, and it was expensive. What do people do today? Do they do back of the envelope capacity planning when they buy something and then pray it is right when they get it? Do they actually monitor performance?
Ash Giddings: People used to be on top of capacity planning, and now it is a bit of a lost art because there’s less and less IBM i skills now on the ground. Most of the time, business partners are engaged to look at problems. So there’s definitely less and less specific skills to look at these issues. I don’t get the feeling that performance is particularly monitored. The power of the machines now is incredible. There used to be many companies knocking on the performance ceiling, telling IBM they need more power. That’s not the case anymore and things aren’t tracked and monitored quite closely as maybe they once were.
TPM: That’s a fair observation. There are very few shops that have more than four Power8 or Power9 cores, much less four sockets, and in a world where there are X86 CPUs with 40 or 64 cores, or with Power10, 16 really fat cores. Do IBM i shops simply throw more hardware at their performance problems? And what a luxury to be able to do that in 2021. . . .
Ash Giddings: With quite a number of clients I’ve spoken to, it is almost finger in the air. They double what they have got when they upgrade, and the cadence of hardware changes has gone up from two to three years to four to five years, and maybe a few more years after that sometimes. If you have all that performance, you don’t need to worry about a little losing a bit here, losing a bit there.
TPM: Well, that Moore’s Law improvement in price/performance should mean that HA is a lot easier to afford, so why isn’t it more pervasive than it is? I have watched that OS/400 and IBM i HA market since the beginning, and in the early days, it was something for the elite because the hardware was crazy expensive compared to what it is today and the HA software difficult to use and expensive, too. So you really had to need it and it wasn’t a surprise to me that only 5 percent of the base two and three decades ago had proper HA. The cost of hardware came down, Maxava and iTera came in and competed strongly with Lakeview and Vision and remote journaling being built into the plumbing of OS/400 leveled the playing field a bit, too. But still only 20 percent of the base has proper HA. This, in a world where ransomware can wreck your systems at any moment.
Ash Giddings: It always surprises me. As I say, HA has been around for a long time and I have been around for so long, too. I got involved in HA in 1994 with MIMIX. And fundamentally, it has not changed an awful lot. And it really surprises me that people don’t have HA. It’s affordable now. What I’ve seen over the last four years is people using disaster recovery in the cloud. They have to dig deep into the pocket for a second datacenter, but not with the cloud. And this has been fueled partly by COVID, I think people couldn’t get people into datacenters to change tapes and do some of those menial tasks that I can’t believe they are still doing. COVID has been a wake-up call for many. So we see lots of activity in and around cloud. But we still have the age-old problem that they can’t afford any downtime. So many of using a combination of the mass storage devices that IBM and Skytap provide to seed the target machine and then you to logical replication to keep that in sync. Because nobody can afford this huge chunk of downtime now. Some companies are running in one cloud and doing logical replication to another cloud vendor, making sure they don’t put all their eggs in one cloud vendors basket.
One of the scariest things, even now, is that there are so few companies that that actually role swap in their HA clusters. I worked with a service provider recently, and we role swapped that client every weekend. It was just part of the tick list, on Sunday morning we are going to role swap. It was a bit of a pain to start with, but after several weeks, it was no big deal and there wasn’t this elephant in the room any more. And when we did have a disaster, it was just role swap.
When you ask people if they role swap, half of them just say no.
TPM: And maybe the other half are lying. . . . Anyway. what are you seeing out there in the field right now?
Ash Giddings: I have only been here for three months, but we are seeing quite a flurry of older, unsupported boxes running OS/400 V5R3, OS/400 V5R4, and IBM i 6.1. Some companies are just stuck on these older versions, because of potential third-party software costs and they haven’t got deep pockets, but they have to do something to protect their machines and we have seen a lot of interest in the Maxava CPR – short for Capture Point Restore – product, which provides some degree of protection and doesn’t cost an arm and a leg.
TPM: How much cheaper Maxava CPR for a typical system compared to going for the full Maxava HA. Is it like one-tenth the price?
Ash Giddings: Yeah, it’s about one-tenth the cost. You can push the data to the cloud or cheaper disk or whatever.
TPM: If I were running a vintage OS/400 or IBM i machine, I would be using something like this for sure and then air-gapping the hell out of it to protect the system against ransomware.
Last question: What can be done to try to get the penetration of HA higher and the penetration of DR to be universal? In my mind, any business-critical system should have DR at the very least, and in a lot of cases should have HA.
Ash Giddings: It comes down to education. One of the most popular webinars I have hosted in years was just reviewing the HA and DR landscape, and the other very popular one was the viability of IBM i in the cloud. There was very little promotion on those and the content was great and we had great people join me. People don’t know, or they haven’t got time to keep up with the options that are available. We can help them figure out which ones are best for them. Just because they put in MIMIX 20 years ago does not mean it is the right thing to do today. This technology has been around for a long time, but it has changed and evolved and there’s still a lot of education to be done.