In our sixth episode of our “I Am A Mainframer Interview Series,” Jeffrey Frey chats with Neale Ferguson, a mainframe expert currently with Sine Nomine, about a variety of mainframe capabilities and use cases. They also take a deep dive into some of the most common perceptions about the mainframe from an enterprise perspective, and how the community can best capture and share the true power of mainframe technologies.
You can listen to the full recording and read the transcript of the interview below.
If you’re a mainframe enthusiast or interested in the space, we invite you to check out our new community forum.
Create a profile and post a selfie with your mainframe system, and you will receive an exclusive “I Am A Mainframer” patch
Jeffrey Frey: Hello, I’m Jeffrey Frey, a retired IBM Fellow and past CTO of IBM’s mainframe platform, and am very much a mainframe enthusiast. Today, I’m very pleased to host the “I Am A Mainframer” conversation series, here with my friend, Neale Ferguson with Sine Nomine. This conversation is sponsored by the Open Mainframe Project, and as a Linux Foundation project, intended to help create a mainframe-focused open source community. It’s also intended to serve as a focal point for development and use of Enterprise Linux in a mainframe computing environment.
The goal of the project is to excite the Linux community around the use of the mainframe, and to foster collaboration across the mainframe community, to develop and exploit shared Linux tool sets, resources, and services in a mainframe environment. In addition, the project seeks to involve the participation of academic institutions, to help in assisting educational programs aimed at developing mainframe Linux engineers and developers of tomorrow.
So Neale, welcome. It’s a pleasure to talk with you today.
Neale Ferguson: Yes, thanks, Jeff. It’s been awhile since I’ve actually seen you in the flesh, as well, it’s nice to put a voice to the text.
Jeffrey Frey: Yeah
Neale Ferguson: So, my apologies up front, for anybody dealing with an Australian accent.
Jeffrey Frey: Yeah, it’s charming. Neale, tell us a little bit about you and your team at Sine Nomine.
Neale Ferguson: I’ve been involved with the mainframe since, 1981 officially, but actually I found some textbooks on teaching myself IBM Assembler back when I was in grade 10, so I suppose I can go back to 1978, but officially 1981 with VM VSE. I worked in Australia until I moved to the US in 1999, where I started doing a lot more Linux work. We had actually been doing a public port of Linux there. I worked with Sterling Software and then I worked with Software AG before starting with Sine Nomine, where I’ve been devoted to doing mainframe and Linux.
ABOUT SINE NOMINE AND EXPERIENCES WITH THE MAINFRAME
The Sine Nomine team is a relatively small consultancy company. We deal with all sorts of things like OpenAFS, which is one of our main open source projects. We have a number of customers running that in production. We do system Enterprise Command Center design and installation everywhere from designing and building, to putting in seats, hooking up OS/390, z/OS, and all sorts of bids to get things working in Enterprise Control Center. My particular area is devoted to the mainframe – z/VM, z/OS, lots of Linux and open source work with our distribution, and our version of CentOS which we call, ClefOS. We have a full CentOS distribution available for people. We do Mono, which is .NET under Linux, which allows you to run a whole lot of Microsoft-type applications. Just recently we’ve been getting into the Docker and OpenShift space, which has been very exciting and I see a lot of potential in the mainframe for those technologies.
Jeffrey Frey: Yeah, that’s very cool. I know that from our history together, Sine Nomine has been literally at the forefront of pushing some of the boundaries, expanding and exploring the role of the mainframe for quite some time, especially in the area of Linux, historically with virtualization. It’s great to see this continue with some of the new technologies. I always say that the mainframe, even though it has its architectural roots back in 1964, is an extremely modern platform, and it’s continually being infused with the most current and modern and advanced technologies. Not only from a hardware standpoint, but now especially with Linux, which is not new on the mainframe, with all of this advanced and current and open software.
Neale Ferguson: That’s for sure. People say, “Look, you’re 1960s technology,” but it’s a testament to its design that it’s been allowed to evolve. It is extraordinary in its breadth, depth, function, uptime, security–all the goodies that you need in an enterprise. The mainframe has and continues to grow. It’s certainly not the stagnant platform that the popular media seems to imply, and it’s been very exciting to demonstrate that by getting things like Linux up and running, and getting all the great software packages that come with the Linux, and now with Docker containers. Seeing what Donna did with that demonstration of working all this stuff together and having it all run in a mainframe and being able to swap over all the resiliency. Everything that a modern organization needs, the mainframe has and continues to be ahead of the curve, frankly.
Jeffrey Frey: Yeah, I share that view with you. More and more people over time are seeing it that way, especially when there’s an ever-increasing need for what we kind of, in somewhat of a cliché way, talk about enterprise qualities of service in terms of availability, data integrity, and especially in the area of security.
So we’ll talk a little bit about more of that in just a moment, but let me ask you, what are some of the recent projects you’ve been working on? What’s exciting for you?
DOCKER, OPENSHIFT AND OTHER PROJECTS
Neale Ferguson: What’s been exciting recently has been Docker and OpenShift. We’ve recently written a white paper, which is available on our site. It describes how you can get the Docker environment, build containers, and then use OpenShift to manage the life-cycle and hook them all together and do all those wonderful things that container technologies have been promising and now are delivering. So, recently, over the last eight months, I think we’ve put out about 50 containers that you can run immediately on Linux on z. It doesn’t matter what distribution you’re running, that’s the other beauty of Docker. As long as the kernel has the right syscall support, then you can run these containers. Whether you’re running Ubuntu, SUSE or RedHat, you can run these containers that will give you MariaDB, Ngnix, etc. Straight out of the bag, you can start being productive without having to go through the procurement and roll-out of a new Linux virtual machine, etc. The containers just make that rolling-out process that much easier.
Couple that with OpenShift, and it’s been an interesting exercise in getting that built on z. It required about three lines of additional code just to recognize the platform, everything else just works because it is just Linux.
Jeffrey Frey: Yeah.
Neale Ferguson: Using OpenShift to then hook these containers together, having them operate, bring them down, have them automatically recover, have the system storage allocated how you want. It’s been a lot of fun with this technology, and I think it’s got quite a bright future.
PORTABILITY, OPENNESS, AND OTHER MAINFRAME CAPABILITIES
Jeffrey Frey: Yeah. That’s great. I think people still do not have appreciation or an awareness of the capabilities of the mainframe in the Linux space. I still get reactions from people that they perceive it to be different in some way, or that is must have proprietary extensions, or that it’s not exactly the same. We’ve always talked about the Linux distributions themselves being as common as any Linux distribution on any platform. I mean, there’s a certain amount of platform binding in any of these platforms, including Intel.
Neale Ferguson: Sure.
Jeffrey Frey: I’m glad you mentioned the fact that these open-source pieces of software, in so many cases we’ve had so many testimonials that you just drop these things on and they work. And not only do the work, they scale better, they perform better, they’re more reliable. We’ll talk about this a little bit in a minute, too, but that is not well understood, and somehow we’ve got to do a better job, and hopefully these kinds of calls will help with that, to spread the word on the value here.
Neale Ferguson: Yeah. A lot of people are still under the impression of the mainframe being a 3270, which I love, but as a user interface for a newbie or those who are used to webs and things like that, it can be a little intimidating. Similarly, they expect, “Oh, if I’m running Linux, then what JCL do I need to run on that?”
Jeffrey Frey: Right.
Neale Ferguson: As much as a fan of JCL as I was back in the ‘1980s, I’m certainly glad I don’t have to deal with it anymore in the Linux world that I’m doing and operating in at the moment.
Jeffrey Frey: Yeah. I know. You’re far more an expert in some of that than I am, but my observation has always been that every platform has a form of that, right?
Neale Ferguson: Yep.
Jeffrey Frey: Every platform has its particular bent on command language, so even in some of those conversations, I think the mainframe gets a bad rap for some of that.
Neale Ferguson: Well, even in the open systems world, the UNIX and Linux world, if you’re a devotee of the Korn shell, then you disparage those who use Bash Shell or Z shell and vice versa. It’s the same old religious arguments that you’ve been involved in, even in the days of NVS versus VM and that type of stuff. So, it’s cultural.
Jeffrey Frey: Yep. The other place that this shows up is when people talk about the need for specific skills for the mainframe. The IBM view has shifted on what mainframe education means, maybe this was because it was dominated by a z/OS environment, where we had to introduce or teach skills that were very specific to use of the mainframe, and I think that’s changing now to be more of a focus on education on systems that are enterprise class. Right? How to make sure that you have data integrity, how you make sure you have absolutely rock-solid disaster recovery, etc.The idiosyncratic nature of the mainframe in a Linux environment has gone away. It’s just Linux.
Neale Ferguson: Yeah. You’re right. It’s going away from teaching people to navigate an ISPF session to business-related functions. The hours, the up-times and that type of stuff that are the important things now.
TECHNICAL FOCUS AREAS OF THE OPEN MAINFRAME PROJECT
Jeffrey Frey: Yeah. So, what has been your involvement with the Open Mainframe Project?
Neale Ferguson: We’ve entered our second of year of membership to the Open Mainframe Project. I’m part of the Technical Steering Committee. We’re still in our younger days, so just sinking into the meat of getting projects and going. I’m involved in the Cloud area with a few others, talking about provisioning and other related things. Some real meat is starting to happen there. My involvement tends to be seeing what’s of interest and what projects are out there. For example, early last year IBM released the Anomaly Detection Engine, which allows you to teach your system about what to expect in log files, and it will suck in a whole lot of log files, and then determines is normal, and then you can run exception reports as new stuff comes in. So, “It looks like something’s weird happening here, here’s an anomaly.” that was released to the Open Mainframe Project, and I added MariaDB support for it there.
So it’s that type of involvement, from the “Is there a software package that may prove useful to the community?” to the bigger tasks of “How do we work in this open, in this Cloud environment, to provision just as easily as everybody else does?”
Jeffrey Frey: Yeah. That’s awesome. You know, I had forgotten about the anomaly detection contribution. When at IBM, I remember when we made the proposed contribution of the anomaly detection stuff to the open community, and it was very interesting in having those conversations. What was great about it is that as we had various meetups, etc, in the community. This fostered a discussion on reliability, availability, and the ability to introduce analytics in a way that was of some real value to businesses. I’m really glad to see that that has been as successful as it has been, and I’m really happy to hear that you’ve been extending that, for MariaDB in particular.
Neale Ferguson: Yeah. Similarly, we built some early blockchain containers that, so far, have been downloaded and used a thousand times in the past year. This is a good indication of the interest in Blockchain. I’ve recently built the latest fabric containers that were made available earlier this year. That’s been an interesting sidelight there, as well. I haven’t gotten into the blockchain as much as I have into OpenShift, but it’s certainly a great technology, particularly for the z. When you’ve got these crypto cards on board and all the goodness that goes around, how secure they are, it’s a very promising technology.
Jeffrey Frey: Yeah. That’s cool. I know you mentioned Donna Dillenberger before. She’s been instrumental in this for z systems as well. Given your work in this area, and as you said, maybe you haven’t spent as much time on blockchain as some of the others, but what do you think? Is blockchain to the point where it really is accepted and will be adopted in a first-class way for enterprises that need that distributed ledger and that kind of a transaction? At first, some people in the mainframe space were worried that either it wasn’t ready for prime time, or people had misinterpreted the use of blockchain to replace the good tried-and-true, two-phase committed hard and reliable transaction systems that z is so well known for.
Neale Ferguson: Yeah. I think it’s one of those technologies that it’s in a superstate of existence at the moment. There’s a lot of unknowns. Those who have pioneered it are progressing with it and doing real things like, I think the noted case study is the people who track diamond shipments, etc. that understand it. A lot of other people think “Well, is blockchain a product on a distant shore?” but not realizing that it’s more than that. It’s an application development environment where you specifically tailor applications based on the technology. Then the other is, as you said, those who are coming from a straight transaction background where, if I run into one thing which is entirely secure, and if I need to talk to anything else, we other methods of doing it, to this really distributed method of cooperation and commerce. I think it’s a technology past the hype stage but on the precipice of either falling into the “Yeah let’s get right into it” to the “I’m not really sure about it.”
What needs to come out of leadership like from the Open Mainframe Project and the vendors like IBM are more case studies and demonstrations and that type of thing, so that people can get a proper feel for it, rather than say, “Blockchain, distributed ledger…Well, what does that really mean?”
Jeffrey Frey: Yeah. Before we leave this topic, Neale, let me ask you about your perspective on Cloud. I have a particular bone to pick about how people misinterpret the essence or the value of Cloud Computing. By this, I mean a service-oriented way of delivering IT, with z. I always considered this as a platform of great strength, with capabilities that were very well aligned with the notion of it as more of a service based way for developing and delivering IT. The prominent attitude on Cloud that enterprises have to move all their workload off-premise somewhere and onto grid-like intel structures. Give me your perspective on Cloud and its value on z, and how we can do better here.
Neale Ferguson: I think one of the driving forces for Cloud is capital expenditure versus operational expenditure. People are seeing large capital items disappear off their ledger, and replaced by just purely operational expenditures. “I need to fire up another several images? Will it come out of my operational budget,” and things like that. To the bean-counters, that’s really appealing. I think one of the things that z suffers from is that they overlooked the price of transactions on z versus non-z platforms, and all they see is the initial capital problem there.
Jeffrey Frey: Yeah.
Neale Ferguson: As in the technology, well, Cloud is just somebody else’s computer. All those things you used to hire your SysAdmins for – somebody else is taking care of it for you there. So, that’s appealing to other people, as well. It’s the “Ask and I shall receive” without all the provisioning overhead from before.
Jeffrey Frey: Right. I don’t want to spend too much time on this but for example, IBM obviously has a business strategy and a business interest in being a service provider, and extending and expanding the global services business in one way or another. Not so much strategic outsourcing, but to really offer a variety of services in the Cloud, where IBM hosts those services and makes them available. Right? But, my interest in z, as a technologist, was always, “Well, how can I introduce the capabilities and the features of a z system that will allow people to build Clouds using that technology?” Right? My focus was always, “Let’s Cloudify the enterprise” rather than try to convince our customers, who have maybe the best experience in running IT shops in the world. Instead of hinting at them that they should outsource the delivery of services to somebody else, even if it’s IBM, is to do more service delivery, be more flexible, adopt new cost models, and have that agility, and introduce that approach to service orientation in their own shops. I don’t think I was ever actually successful in getting people to kind of see it that way, or at least enough people to see it that way.
Neale Ferguson: Yeah, I agree. Frankly, the Cloud model for somebody else’s computer is appealing and right in a lot of circumstances. The hybrid model is the type of thing that we should be enabling on the mainframe, and making that simple. It’s not easy.
Jeffrey Frey: Yeah, and we’re going to get to this with the thing I want to talk about next. Some of the challenges that we face in IT, and especially in mainframe space, are that some of those challenges are not technical in nature. They’re either perception, attitudes and either a lack of appreciation or an unwillingness to change processes or business models. I think that applies to Cloud in particular. If in fact, what’s preventing an enterprise from hosting services, or getting services delivered on somebody else’s infrastructure, means is that they haven’t taken a look at taking their own IT environments, and potentially making those IT environments more of a profit center, rather than just a cost center. Right?
I think it’s an apples and oranges comparison, because the way IT shops are managed, in on-premise customer location, is not the same business model as services that are hosted in what people think of as traditional Cloud.
Neale Ferguson: Yeah, I agree. Yeah.
Jeffrey Frey: So, not at all a technical set of problems. It’s more a set of issues related to how enterprises perceive, treat, manage and operate their own IT environment.
Neale Ferguson: I agree. If it was purely a technology issue, then it would be relatively straightforward, but there’s so many cultural and inertia-based things that make the real difficulties in moving a particular direction.
MAINFRAME CHALLENGES
Jeffrey Frey: Yeah. That leads us to the next question that I have for you, which is: What do you consider the biggest challenges for the mainframe, going forward? Are they the same, are they newer than what we’ve kind of identified in the past? How can we make the mainframe more accessible and more valuable to people?
Neale Ferguson: I think a lot of the problems are the same that we’ve had over the last umpteen years, that have just, in different guises now. We still face the us-and-them situations in workplaces, the silos that build up. People can bring up the old great business cases for doing something on z. Those business cases can be then taken by the others, who then can either warp the figures by doing a cost of acquisition over shorter periods and that type of thing.
Jeffrey Frey: Yeah.
Neale Ferguson: You know, “You’re going to need to build another building if you want to do this here, or “ Well, we’ll just factor it in because we might do it anyway, so we don’t count it and we then compare it against your mainframe model” and, “Oh, look, we’re so much better.” We’ve got still that type of problem with silos, of people who graduated in the non-mainframe environment, who are now reaching the sea level and not having that background. It’s hard to educate there because they’re in such a position where they are not being paid to be educated but to make decisions.
So, culturally, I think the mainframe issues are the worst and the hardest to deal with. The technology issues, improvements in functionality and reliability and all those other things that the z’s been doing over the last few years. If you’re evaluating on that basis alone, then it’s not much of a hard case to make. It’s just all those other peripherals, for want of a better phrase. You either stop it from being put in the door, or have pressure on it while it is in the door to get it out of the door.
Jeffrey Frey: Yeah. Yeah, there’s probably an entire couple hours we could spend on this topic of challenges
Neale Ferguson: Yeah.
Jeffrey Frey: You know, I know that you have kind of the same view I have on perception and apples-to-apples comparison on business case and value and cost in particular. You know as well as I do, we’ve run into environments or enterprises where “the cost of the cafeteria” is in the mainframe budget somehow.
Neale Ferguson: Yes, yes.
Jeffrey Frey: Yeah. I get the feeling that some of some of the folks that actually love z, and are running successful z-shops, can actually in some cases be part of the problem. What I mean by that is that, you know, it’s probably no secret to anybody that one of the reasons, for example, that z has a reputation for being stable and reliable and available is because its managed well, right? The people who manage these systems in large enterprises are very skilled, they’re very disciplined. They’re very process-oriented. They have extremely well-honed processes for change management, etc. which maintain order and stability in these systems.
The downside of that is that they’re often perceived, and maybe rightly so, as being protective of the environment, which leads to not having the environment as accessible to people or to developers as other platforms
Neale Ferguson: That’s certainly a good sentiment there. A lot of people joke that DevOps is put into production first then test second. If it’s used properly, and z can fit into this new way of doing things quite nicely, where you can quickly fire up your apps and demonstrate their value and promote them in through the systems, though have a well-managed life cycle, like OpenShift. It brings that similar discipline, where we’re used to, in the z environment, of being cautious but courageous. It seems far more accessible that people can spin up and get things done, like ad hoc projects, and just good enough to get things to the stage where they can be brought on board, and they can be brought on board on the z finally, rather than somebody having to fire up a little VMware instance and they can demonstrate it there and then promote it into production through the web cycle, and then make the conclusions there.
Jeffrey Frey: Yeah. Those are great insights.
Neale Ferguson: The other thing we like is a lot of work here at Sine Nomine is not just looking at the future technologies, but helping to exploit things, and even to the point of things like wonderful technology called NJE. With the mainframe operating systems are able to transfer files. You know, I see a lot of our clients using all sorts of bizarre FTP scripts to try and achieve the same functionality, so we actually created an NJE, and since it works on Linux and Solaris and that ports a pier to z/OS, and VM and VSE and, like using even over SSL on the top of stuff.
Jeffrey Frey: Yeah.
Neale Ferguson: It’s an old technology but it’s a really, really powerful one if it’s used properly. If you can integrate your mainframe and your distributor systems in using technologies like that, I think that’s always a positive, too. It makes the mainframe visible, “Hey, the mainframe is serving this capability. It’s reliable, it’s secure, etc.” I think those type of small things help. If you can exploit the strengths that have been there for a number of years, is one of those good news stories that will help the mainframe’s cause.
Jeffrey Frey: Yeah, that’s a great point, because some of those capabilities and functions are not very glamorous. Right?
Neale Ferguson: It’s the plumbing. If you get the plumbing right, you don’t have to get the plunger out.
Jeffrey Frey: Yeah. That’s exactly right. Another great example of that is people who can take advantage of the enormous I/O bandwidth and performance and sustained I/O rates out of a z system for doing data backup and recovery, and transformations and movement from one platform to another. What we’d like to have happen, of course, is that we find technologies and approaches to data management where there’s not so much movement of data around, because there’s one hell of a lot of capacity spent just on moving data from point A to point B, and doing various types of transforms. Given that that data has to be moved around so much in an IT environment, and it has to be backed up and copied etc, z has great facilities for doing that kind of stuff, and it’s not very glamorous, but it is plumbing. It’s necessary, and doing it on z is cost-effective.
Neale Ferguson: Yes. Well, one of the problems that z has in that respect, though, is a lot of these new technologies get built almost as a prototype on a smaller platform. When a company gets to the stage of rolling it out, they start finding the scalability issues.Then they resort to various technologies like, MongoDB has Sharding. Well, the difference is, if co-developed on z, they would have discovered that you don’t really need Sharding if you have the right infrastructure and things like that. We’re trying to fit things into the packages that run on z that don’t need to be there, because the hardware can take care of things and the like.
Jeffrey Frey: Yeah. There are lots of examples of that. Maybe if things were different and turned around, where a lot of these technologies were first developed and introduced on z, with its capabilities to scale, we would still have designs that would, in those cases, need to scale down effectively, anyway. So, maybe there really is no single design point.
Neale Ferguson: That’s right.
Jeffrey Frey: But there is a presumption, right? As you’re pointing out, there’s a presumption in the design of a lot of this software, that in order to get scalability you have to implement very creative and innovative ways of doing some of that stuff, that are really not necessary on z.
Neale Ferguson: That’s true. I mean, cluster for good reasons or bad reasons, cluster because you just haven’t got the capacity. One of the things we provide is, for the z customers because Red Hat doesn’t offer the service, is the higher viability option. So you’re using the GFS-2 file system to be able to have multiple modes with concurrent rewrite access. There are good business cases for where that technology’s good, and z fits right in there with it. But then you’ve got the other places where clustering is there as a means of scaling or getting around some other bottleneck that ought not to be there.
Jeffrey Frey: Yeah.
Neale Ferguson: So, I think that’s, you know, leads me onto just saying that one of the nice developments has been the LinuxONE Community Cloud system, where people are developing this type of stuff from scratch. They can work concurrently on z as well and they do it on their Intel boxes, and see what the different characteristics are and understand how they can take advantage of both environments and the like. So, that’s been a good innovation. My pet peeve is always there’s never been a hobby license for z software.
Neale Ferguson: But, that would be the next step beyond the LinuxONE Community Cloud, as far as I’m concerned, if you want to continue this outreach and for new workloads and new people getting on the platform.
Jeffrey Frey: Yeah.
Neale Ferguson: I mean, one of the things that we’ve done, I’ve been doing for a number of years, is we’ve got the Mono project, which allows you to run dot-net applications, because it wouldn’t be nice to bring in that type of workload into a z environment and let them run their dot-net apps other stuff. Again, you get the plumbing right and make it available, then there’s potential for good things to happen. There’s a whole lot of surrounding stuff that needs to happen as well such as having a community system, having hobby licenses, getting involved from day one in developing packages on z, are some of the things that need promotion.
THE FUTURE OF THE MAINFRAME
Jeffrey Frey: Yeah. I couldn’t agree more. That kind of is a good segway to maybe the final thing I wanted to ask you about, which is, especially within the context of the Open Mainframe Project: What would you like to see that community engaged in, and what would you like to see come out of that community in the future? And maybe some of what we just talked about, maybe the OMP could help with.
Neale Ferguson: Yeah. I think this outreach, its visibility to the outside world where participation in bounty type projects, participation in mastering the mainframe, the Linux community system is excellent. Can we extend that beyond just Linux space to our people to extend, so they can develop multi-tier applications easily, without having to run through all sorts of hoops or indulge in all sorts of acquisition of hardware, etc? I think the Open Mainframe Project can advocate for that type of thing and work through those technical and non-technical issues to make it happen.
Jeffrey Frey: Yeah. Well, I’ve eaten up more of your time than allotted. Neale, Let me just thank you once again for taking the time and sharing your thoughts with us. I know it’s valuable for me, and hopefully those who listen to this podcast will really enjoy the conversation, so thank you very much.
