This week, the Terminal Talk guys sat down with Steven, LinuxONE Global Sales Leader to talk about his favorite system and how this exciting offering fits into the broader ecosystem.
You can listen to the podcast in audio form by searching for “Terminal Talk” on your favorite Podcast platform, or by heading to the episode web page here.
Frank: We have with us today Steven, the LinuxONE Global Sales Leader. That’s something we don’t normally say on this show.
Jeff: Yes very fancy…
Steven: Thank you guys. Thank for having me on.
Frank: Could you start by saying what the heck is LinuxONE?
Steven: Okay so LinuxONE is an IBM brand in the systems part of IBM that we’ve launched back in August 2015. It’s a enterprise limit server from performance availability security and scalability. So we’re positioning that brand to specific clients who are looking to do Linux workloads and turn those dials in the four areas I talked about performance availability security and scalability up to eleven. So if you’ll permit this Spinal Tap reference.
Steven: And so in this – with this crew it would be okay. But so it’s really positioning that for those workloads where clients are really honed in on the ticket with one.
Jeff: And that area would typically be like where I’ve seen it strongly positioned as the bank finance where they typically have a strong Linux footprint already. And you Tweeted about something earlier this week about Temenos. Is that being like a good measurement or a hallmark of where it will be a good fit? Can you also talk about that?
Steven: Yes exactly. So I think some of the verticals where we see the strongest requirements for a high-end limits server as a traditional industries that are focused on scalability. So hundreds of clients or interactions or transactions, you know, banking would be a classic example of that retail problem that’s (the healthcare sector). So those would all be sort of the really interesting verticals for us. And as you say I recently Tweeted and blogged about what we’ve been doing with Temenos.
So Temenos for those who don’t know are probably one of the leading core and keen software providers. So if you want to buy a commercial off the shelf banking application.
Jeff: Which I might do, you know.
Steven: Yes, yes I mean I wouldn’t recommend you run it on your laptop.
Frank: Jeff’s Bank.
Steven: You’d be surprised by the amount of new challenger banks. You know, so we’re not talking the Citi Group’s here, we’re not talking the Barclay’s Banks we’re talking about new banking organizations. Temenos and their T24 application would be a perfect fit. I was there at their conference a few weeks back in Dublin and I was surprised that they’ve got 3,000 clients in 150 countries.
Steven: So that gives you a scale and a perspective. So the reason why I was Tweeting and blogging about it is because we’ve just run a benchmark down in our Montpelier Benchmark Center. And some of the points there are just fantastic from a benchmarking point of view 12 X faster for open or batch processing. So if you look into reconcile current accounts as a bank as an overnight batch process doing that in an hour versus 12 hours is obviously material to your bank.
Most of the other benchmarks were 4 to 5 X faster bigger quicker, you know, whatever the metric was the comparable x86 platforms. So I think if you’re a sort of Tier 2 3 4 bank much larger of applications Temenos will be on your short list. If it’s on your short list then we should be the platform to support them.
Jeff: And I’m – I’ve grown to be wary of 4, 5, 10 X claim why do you think it is faster? Is consolidate more and play smarter with your resources ?
Steven: Yes so there’s an element of that. So the Temenos back end is typically an Oracle back end. So that’s been a workload just from the get go with LinuxONE. So that’s just a data serving at scale platform like typically we see a 10 to 1 consolidation for Oracle databases. So if you’re looking to buy Temenos its license is probably transaction based this isn’t going to be material to you at all. But you’re still going to have to look at getting Oracle licenses to support that back-end database on a Oracle database because at $47,500 thousand for a per core license and then you’ll spend $15,000 on RAC and $15,000 per core for their encryption software the costs add up pretty quickly on your back end.
So we’ve got the sort of classic Oracle consolidation back end story. And then it’s just these I/O heavy applications. So from our point of view just engineering in the box we’ve got a dedicated I/O you don’t have that in a powerful x86 environment we’ve got a faster processor. We’ve got more cache, bigger cache. You know, it’s deep engineering in the box means we can get to that…
Frank: So I just want to pull in that thread a little bit because you talk about things like encryption. So I have the capability to run some encryption.
Steven: So we can obviously take a lot of the capabilities that are coming from other IBM server and apply those to what we’re doing in a Linux concept. So hardware security monitors encryption on the chip and the new technology we’re launching that we’re starting to talk about now at most the Think conference around Secure Servers Containers.
Jeff: He’s good.
Steven: All of those technologies come to bare in a Linux environment. So FIPS 140-2 level 4 deep techy sort of compliant security, type stuff pure back and you’re looking at payment card industry data regulation the PCI-DSS guys and looking at that and starting to mandate at the banks. If you’re comparing whether you put it on the Cloud on whether you put it on prem most Clouds including the big three or four are all FIPS 140-2.
Steven: You know, so all of our security value propositions just built on some pretty detailed pretty tech low levels stuff that blows my mind but, you know, real engineering that comes out of the guys here in Poughkeepsie.
Frank: So again this is kind of important in a world that, you know…
Jeff: In a world.
Frank: …is looking…
Jeff: …in a world that’s…
Frank: …you know, in a world that’s…
Steven: You should be a voice-over artist
Frank: …but in a world where it is a big deal is this kind of a value add of the product or are customers saying that makes sense to me because I need to be able to support that?
Steven: So we just based on (COL) and CTO and to FIPS and it was fantastic for me to be involved in a bunch of those. And really what we were doing in those was grow testing our value proposition. And so in those calls we’d spend half an hour understanding that was challenging and our journey’s Cloud. And then we spent the second half of those calls road testing and the . What out of the value position works?
And out of the 30 calls I was involved in literally two are man and a woman on those calls every one of them the minute we started talking about security. So, you know, obviously when you’re out there talking to customers they’re polite enough to let you tell them about how great your stuff is but they found it very difficult to mask when they’re genuinely interested. And so you – it appears the (wholeness) of the conversation and the level of questioning and the amount of interaction instantly changed as soon as we started talking about security.
I read recently that the average length of tenure for a CIO is half that of a CEO. I think you can manage – you could measure in days the length of a CIO after a major security breach.
Jeff: …oh yes.
Steven: CEOs got to throw someone under the bus and got to declare that we’re moving forward with a new team and the fall guy in that situation is going to be the CIO. So I think that’s in the back of the mind of all of these CIOs. So you mentioned something that’s going to help them with their posture help them with compliance move from away from people and process and move to technology that stops the rest. That conversation just needs to be changed.
Frank: And apart from other platforms right LinuxONE isn’t saying just hey that security thing is the thing that differentiates us right. I mean the fact that to do this is cool and an important part to a lot of CIOs but you’re not going around saying LinuxONE our basic computing plat .
Steven: I think as we’ve looked at how we’re positioning the platform over the last three years I think we’re lead on maybe the other three of those value props the sort of performance availability scale maybe bring TCO savings into that equation. Those were I kind of lead with messages. I think as the engineering teams and element teams have added out our security story with things like secure service container we are increasingly leading with security because Linux can run on everything. We were joking about it before we started everything from a raspberry pie through an Android phone.
The guys at the Linux Foundation talk about this, you know, everything from a raspberry pie a phone a TV a screen in the back of a , you know, right up to (one of these) computers.Steven: So how do we capture a workload for something that runs on everything from a raspberry pie right through to a LinuxONE. So…
Frank: …right and so, you know, I’m probably not going to run my Icecast server on a LinuxONE right. So how do you determine or how do you help a client determine this is a good time this is a good workload or this is…
Frank: …or workload or my business really needs one of them?
Steven: …so I mean I tend to be pragmatic in that conversation so I think maybe previous conversation have been. We can sweep the floor with all of your Linux environments in technically we can you can run a Web server on LinuxONE and it’ll run great and you want to put some and I (imagine it’ll) – it’ll run like a scolded cat. You know, it’s like great.
Jeff: I think it would run better on like 4 or 5.
Steven: Yes I mean if we could talk afterward yes. We’ve got some videos and we’ll talk. But no I mean all joking aside I think with where clients are with Cloud where Public Cloud are moving to a Cloud for the on-premise deployments I think you’ve got to be pragmatic. So while you tend to have that conversation with a client this way. What are the 10% of workloads that keep you up at night? Not go in there with big Bravado and say I’ll sweep the floor of all of your x86 environments and rip out 10,000 servers.
It’s where’s the credit card data held what are the five ten 15 servers that if you get hacked in the CIO you lose your job.
Steven: I mean it’s around how you sprain the conversation together. I think I was always told people for three reasons make money save money not get fired.
Frank: And probably not in that order.
Steven: And that’s the point everybody’s out there having a – my box can save you money my widget can save you money. People have got tired of that. My box will save you 30% story because the benefits are hard to measure they’re typically 12 18 months down the track. You’ve got to have a successful project to get there. Talking to a CIO or a CTO and saying what’s the environment that you’re most worried about. And we’d have that conversation to say, you know, where’s the patient record data? Where’s the credit card data? Where’s the the ideas-based data?
Where’s the, you know, where’s that mailing list that gets you on front page at the Wall Street or if you get hacked. You see the conversation change in the client’s mind and they go oh yes, yes, that’s that so it’s the service over there or recall or post or MongoDB. If that goes down on .
Frank: So yes.
Steven: So that go and have that conversation first and then we might be doing a consolidation and savings some money afterwards but let’s make around the biggest pain point.
Frank: So the way you’re talking about it because before you were mentioning positioning LinuxONE x86 but increasingly it sounds like you’re positioning it against their (file) .
Steven: Yes I think I’ve read lots of research and from chatting to a bunch of clients I think we’re starting to see that and run to the Cloud sort of temper a little bit. I think people are being more pragmatic about hybrid on and off premise. (And) this is my view Steven’ view of where we’ll be five years from now. I think 80% of science workload will sit in the Cloud and 20% will sit on premise behind the .
You know, if you want to run a Web server these days and run an agile environment and you want to go and run a mobile app, you know, unless there’s really good reasons the security or availability performance scalability and data residency point of view on premise it’s going to be a tough argument to win, you know, let me run your Web servers on premise. That’s (ship) sale. And that’s probably the bulk of somebody’s x86 server environment volume of servers that’s the that stuff’s gone.
That’s the stuff you should be putting on the IBM Cloud. The 20% is where we need to . And then within that 20% what’s the stuff that’s get if it gets hacked you get fired. That’s the probably the 1% of their overall install base which is where we should start the conversation.
Frank: Can you talk a little bit about secure service container on LinuxONE? Could you kind of describe that a little bit?
Steven: So it’s deeply techy stuff, you know, and Frank that’s not my area of expertise but I’ll tell you a story. So if (Mark Quigley) listens story so there’s the copyright infringement (out of the way it’s out of the way). So you’re a service provider you’re providing service to a number of different clients and the FBI turn up and they’d say we’d like the data from that particular client and they’ve got a warrant. Now you can go down and ask your Sys Admin he’s got elevated access rights who you’ve given those rights to and you’ve locked him down through people and process and procedures.
But technically that Sys Admin can go and get that data download it out of the virtual machine and serve that warrant. From a LinuxONE perspective with secure service containers what we’re doing is putting that data in what we call secure and (parts) of container. It’s a bad word now with . What it effectively does is puts that environment in an encrypted (enclave). And what that means is that when that person with those elevated access rights go to get the data all they get back is encrypted data.
So you can’t serve the warrant. You can’t provide that data to the FBI. Now I’m not saying and fantastic guys and girls in the FBI or.
Frank: Certainly not.
Steven: Yes but I tell the story in that way to give you a context for how you would use that. So imagine you’re still that same service provider and the person with your is like me and he’s got four young children. You kidnap those young children I will do anything as a father to get them back. You’ve now socially engineered me and people of process aren’t going to stop you…
Steven: …in that data to the people that have got your kids hostage. So you’re able to get to the data and you’re able to get it out. When you look at cybersecurity threats about 30% to (40%) are by employee. Do you know the full background of that (VM) administrator last week? Do you know the full background of that contractor who you’ve just given access rights?
Yes you’ve got a process yes you but yes you’ve got a way to do it but you haven’t got a technical engineering way to ensure they can’t get the data. So secure service containers fantastic piece of technology other people will do a better job of explaining it technically but for me that’s the use case not just putting people in process around data security actually engineering the of the data that’s the value proposition.
Frank: It’s probably wrong to think about I’m a hardware I think about the box itself, you know, I’m the…
Steven: The support groups for that don’t work you’ll be fine we’ve got you. We’ll look after you.
Jeff: I’m looking at the box and I’m looking at the components and stuff like that and I’m thinking about it wrong because I’m thinking about it a certain way. When the LinuxONE movement is more about the whole ecosystem and the software that comes with it that’s a whole part of the Web page when you go to look at it . Can you talk a little bit about like the vetting and the porting and all of the software that, you know, comes into that?
Steven: So I mean selling a box the part of the value chain is the box we’re selling so. For me I – what I spend time is I that translation pointing to where some hardware engineering has done some amazing work over the last five years what does that mean where the stack? So let me sort of tell another story. I’ve told you this …
Jeff: I like stories.
Steven: … one. So you want to run MongoDB and six out of the ten are the fastest growing databases are Open Source databases.
Steven: So if you’re deploying a new app you’re probably not going to be doing your Oracle you’re probably not going to be doing and (sorry). I’ve probably offended some DB2 people along the way there but you’re going to be picking a a Mongo or a (Scarlet). You know, one of the Open Source products. Six out of ten of those fastest growing databases were Open Source databases and that kind of . So you’re picking them MongoDB and you want to run that on an environment. Well we’ll then go and have that we could say well you can run that on lots of things. You can run it on the Public Cloud you can run in on Commodity x86 you can run it on . So say you want to run a big banking application and you focused on performance.
So the number one reason why people change their banking is they’ll change their bank is because of the mobile application. So I bank with HSB. See my new branch is two and a half hours away. I don’t go into a branch anymore but there app is fantastic and I’m a loyal customer. So you deploy MongoDB availability performance scalability type reasons and you want to translate from the hardware techy stuff down in the box to what that means when you’re deploying MongoDB.
Steven: You can deploy MongoDB on anything, so why deploy it on a LinuxONE? So a number of different reasons, for me it was because of cache. So up to four levels of cache in our box was important. It means you can do just more through the particular processor. So when you approach transactions at scale it’s important to have memory. The amount of total memory important. Well if the data that is across multiple x86 environments and the memory in that server fails you got to do a fail over. We’ve taken that scenario out.
I’ve chatted with a client recently who’s got three copies of the database in their own x86 environment. We just turned around and said well we don’t need that level for this application, because we’ve done it down in the hardware so we’re working with that customer to re-engineer their application for two copies of the data not three copies . You look at just a role of kit performance. You look at the I/O system. You know, all of these very techy things, who i am not the right person to talk to you about translate.
So what does that mean? So you look at the MongoDB benchmark with a 17TB Mongo database without the need to (shard). It means you don’t need lots and lots of small servers and having to run around work out where the data is you’ve got it in one big DB instance, that runs like a scolded cat. You know, so if you’re doing a banking application then it’s one MongoDB database and you don’t need to shard it but you’d have to shard it from an x86 point of view..
Steven: So from a solutions architect point of view that stuff makes sense. As I say I’m probably the worst person to have a conversation. But that’s the triumph of architectural battle which are on a win with this box.
Jeff: that again is something that is so fully baked into the distributed Cloud type model it just disappears when it’s on a hardware.
Steven: Yes so those for hardware find you when…
Frank: And it’s literally part of it…
Steven: … yes. And what we can change that dialogue and this client had talked about where we’re talking about three copies of the data to two it just took a third of the number of calls that that would not having to do any other smart stuff we were able to eliminate a third of the . We just didn’t need that compute capacity because they were buying that compute capacity to mask of at the memory level for those guys.
Frank: …right and, you know, the whole concept of consistency can (disappear) and they don’t have to worry about all that (shard) up right…
Steven: …so those guys, you know, Docker containers virtual machines all of this technology is fundamentally there so that workloads are staple less so they can move around on an x that is designed with failures built into it. When you take those failures out I’m not saying not to run Docker and I’m not saying not to run virtual machines just use them for other reasons as opposed to being a mask for hardware failure.
Frank: Yes I wanted to make sure that you finish up with Jeff. We do support Dockers…
Steven: Dockers mess with the …
Frank: …we do that…
Steven: …pick your orchestration a couple of choice we’ve got you covered but yes you can use those for orchestration and value add reasons as opposed to as a as a mechanism to cover you from memory failures.
Frank: …so besides the Brooklyn accent that you have.
Steven: It’s a giveaway.
Frank: Yes you do other stuff right? You have opened project can you talk a little bit about that?
Steven: Yes sure I have the pleasure of working with the Linux Foundation back in the early part of 2015 with of what’s called the open mainframe project. So the way the Linux Foundation is kind of structured now is more around collaborative projects. So the is Hyperledger right now so that’s a collaborative project (structure) I think they’re (up and over) 350 members. And what they do is provide a kind of central hub as you – if you will in the community for anybody who wants to collaborate around Blockchain .
So Hyperledger project the underlying technology is Blockchain. structure open mainframe project underlined technology is the mainframe and LinuxONE platform. So if your computer associates you’re your Velocity software your . If you got your business built around all your academic institution pushing this type of technology. So there was a community already around this stuff that have informally kind of gathered around the technology. The open mainframe their Hub.
So we’ve got a structure that’s similarly exactly the same structure as you would have for Hyperledger or ODPi or (R) or, you know, any of these other Open Source projects. My favorite is the open drone project. (That’s where a community is) gathered around that Linux as operating (systems) drones. I’ve got to find a way to interact with that one…
Frank: Yes that’s a …
Steven: …but well automotive Linux which is how they get together to push Linux as an operating system for cars. But joking but those it’s a gathering place in the community. So we’ve got academic members we’ve got students (coding) for us. We’ve got 12 done this summer and over an internship project. We’ve run hackathons, you know, with Open Linux and Open Source on this computer.
Jeff: And how can, you know, this is a good chance for you to plug your – so our literally dozen of listeners can find out where …
Frank: (of them)…
Jeff: …yes …
Steven: Well we’re going to try and fix the dozen of listeners…
Jeff: But, you know, where can somebody find out more information about that and your efforts?
Steven: So you can either go to Linux Foundation.org and then go into the Collaborative Program page and you’ll find us there. Or you can go directly to us as openmainframeproject.org we’re on Slack we’re on Discourse we’re pretty easy to engage with and you can join our mailing list. As an individual it’s free to be a member so you get on and we tell you about what’s going on on the project and give you ways to get involved. There’s coding projects you can get involved in maintained under the project.
So you can as a developer just get stuck in and act and get stuff done. If you’re a corporate member and Linux is of the mainframe platform…
Frank: Jeff’s bank.
Steven: …Jeff’s bank and you want to be a member we’ve got a couple of clients who are members of the project now and we’ve got to get involved all software vendors and hardware providers to get involved as well so.
Steven: Also mainframeproject.org.
Frank: …and you got to make sure you say bank Jeff’s bank.
Jeff: Oh, oh…
Jeff: …that’s the on my plan.
Frank: Yes. So anyway this has been great Steven we really appreciate you coming over. Maybe we should have one of those educational type to be on the show sometimes though.
Steven: Harry Williams from Marist would be perfect.
Jeff: That’s a fantastic idea.
Frank: Yes if only we…
Steven: He’d be awesome…
Frank: …could get him to come .
Steven: …if only you could get him to come two miles down the road.
Frank: Yes. Well again thank you very much and…
Jeff: (We’ll) have to Skype him (then.
Steven: No we’ll take the studio to him.
Frank: We’ll work on that.
Jeff: In the meantime old man Charlie .
Charlie Lawrence:You’ve been listing to Terminal Talk with Frank and Jeff. For questions or and and on future episode direct all correspondence to firstname.lastname@example.org. That’s email@example.com. Until the next time I’m Charlie Lawrence signing off.