Episode #14 IBM AI OPS, the AI system management system that learns, protects, recommends, and fi...

Thomas Anglero talks with IBM CTO Mike Goretski about AIOps and multi-cloud management. IBM's system gives you visibility across fragmented cloud environments, cuts costs by automating routine tasks, and integrates with your existing tools like ServiceNow and Red Hat.
Chapters
Computers now have primary control of critical vehicle functions. Welcome to the IBM Keep It Simple Technology Podcast, where each episode we discuss technology, but in a simple way. We keep the explanations simple, but what we spend our time on is emphasizing what this technology means to you and how you can best utilize this technology for you and your business.
Welcome to the IBM Keep It Simple Technology Podcast. I'm your host, Thomas Anglero. Once again, I have another amazing guest who we've been talking off air, and this is going to be a wonderful podcast. Ladies and gentlemen, I'd like you to meet my new best friend in life, Mike Goretski. He is the IBM CTO of the Nordic Cloud Acceleration and Transformation. Mike, did I do your last name right?
Oh yeah, Thomas, thank you very much.
Mike and I were saying we both have hard last names to pronounce. So I promised him to do his last name in his native tongue or close to it. So that was fantastic. Cool. Thanks, Mike. Listen, Mike, we're going to jump right into it. So a few weeks ago, IBM at this first all-digital yearly conference, IBM Think 2020, year 2020, it announced what I truly believe was probably the most important announcement, coolest announcement, whatever adjective you want to use. And it was called AI Ops. And I've been so much looking forward to doing this podcast because every single person listening to this podcast, even if they run a dry cleaner, even if they run a hair salon, are going to love the concept of AIOps and will want to use what AIOps is. Am I right? Do you love it that much too?
Absolutely so. I mean, this is the melody of future, right? So I love it.
So Mike, let's, let's go right into it. Please tell everybody, and in simple terms as you can, the big announcement on Watson AI— not even Watson, it's not Watson AIOps, it's just AIOps, correct?
Oh yeah, AIOps in general.
AIOps in general. Please explain to us, the floor is yours, what is AIOps?
Sure, thanks Thomas, and thanks everyone for tuning in here. So AIOps Basically, in short, and let me just put an explanation into the acronym here, artificial intelligence in operations, right? Or if you so will, in IT operations, right? That's what it stands for. So what does it mean and what does it do and how come we actually focused on this and identified it as one of the focus areas going forward, melody of future, as I mentioned, right? So we basically started looking into this based on a number of customer pain points, right? If you start at that end, And there are many of them around, but I just focus on 4 most significant ones. One of them being fragmented visibility, right? So the customers in this hybrid multi-cloud world that we are in, that we have already entered, right, will have different sets of workloads running on different sets of infrastructures, or if you so will, vendors, environments. And so all of those will have to have some sort of an ITSM, IT service management, capability around them, right, wrapped around them. And at the moment, there are very few or none who actually can offer visibility to other vendors or other types of infrastructures, IT service management. So basically, you wind up with a number of different tools in a number of different environments, right? So that's one of the pain points. Having said that, of course, to maintain, manage, and actually keep all those tools and processes alive and actually adapt to these, you need to have skills, you need to have people who are actually up to speed to maintain and cater for these capabilities on different platforms, on different technologies. And that, of course, drives cost and complexity. So that's another sort of pain point here. Of course, the risk exposure is the third area, and that's due to the reactive capabilities of these tools versus being, having to be proactive in complex environments, right? So that's another sort of driver for artificial intelligence to be introduced here. And last but not least is the inability to harness cognitive insights, right, by leveraging and having knowledge about different sets of data in different sets of environments, right? So that is basically why we came to this conclusion about AIOps and its sort of unified capability to be offered as an umbrella over different environments in a hybrid multi-cloud setup?
Okay, Mike, so now let me simplify what you just said, if I can. Sure. So I think the key word that you said to simplify everything really is that this is— we're talking about IT service management. Am I correct?
Absolutely so.
Exactly. And that's why I said even a person who owns a dry cleaners or a hair salon, this is relevant to them and important to them because this is the— well, if you have a business, no matter what the business is, if you even have a cafe, you want to make sure that your company's website is up and running and is working as best as possible. If you have a more complex business with databases running different services and things like that, it's even more important to you that your IT service management is at its best. And that's what is AIOps. And what you mentioned was that it is— I'm looking at some of my notes here. One is the 24/7 management. I added the 24/7, but the constant management of the system utilizing the benefits, the beauty, the hype around artificial intelligence, its cognitive capability to learn, take in data from all over the world, and to run your IT system best. Is this— am I saying it in a very simple way? Is that correct?
Absolutely. And just to add into that, from different various different sources in a hybrid multi-cloud setup, right? And if you have your data or systems across many different vendors, infrastructure, or data centers, and in your own, of course, in addition, then of course you need to have something centric in the middle of everything to be able to have a sort of unified view into all these environments, right?
So what do you mean by centric? Let's simplify that. You mean one system, like a one master system to watch all the other subsystems?
Exactly so, or mother of all other systems, right? ITSM systems, if you saw this.
Exactly. And this is— and now let's emphasize it for people that we're talking about a software. We're not talking about a person because prior to the announcement that IBM made a few weeks ago at IBM Think 2020, usually you had people and these people were monitoring your IT system, correct? And I guess in phase 1 of the IT world, we hired people in our company and we had our IT manager and our IT manager had his team or he or she monitored themselves. Then, the world in phase 2 then outsourced that to save money, right? But now, what we're talking about is this next phase where actually AI software, in this case is IBM's AIOps, is so intelligent, it has learned so much over the past decades that the software is now doing the work of the people but managing it, what you're saying, with almost infinite scalability, looking at all the different systems that you have. So if your company is a small startup, no problem. If it grows to a multinational like IBM in 200 countries around the world, No problem. Is this correct?
Yeah, that's absolutely right. And so, and also to add into that, Thomas, it's very interesting. It's simply not about augmentation of people, right? Right. Exactly. Providing the capability or additional capabilities which would take humans days, minutes, hours, or whatever you so will, right, to actually accomplish. So instead of getting your, composing your sunrise reports, utilizing a couple of hours in the morning, you would actually receive a report as you are driving to work, right, through an app. Basically, are the systems up? Have I had any troubles, issues? Were there any alerts, right? And then it gives you enough time to make the right decision in a timely manner by the time you get to the office, right?
Exactly. And I think you nailed an important thing, and I hope I'm— let's emphasize that one more time. AI is not here to replace us. It's here to assist us and supplement us, right? That is such a strong point. I want to make sure everybody listening hears that, that this is not just freeing up your time, but it's giving you the world's best knowledge and information at all times so that you could take your company to the next level. That's why I'm super excited about this because I'm a startup guy, right? I love startups. I've done startups. But I don't have the money to hire the world's best IT service managing person. But I do have the money to have this software, this AI system. This is actually the fulfillment of AI, isn't it?
It is indeed.
Do you have some examples? Have you recently worked on a project that utilized AIOps?
Absolutely so. It was in a very exciting project, a proof of concept actually, which is turning out to be put in production. Pretty soon here with one of the largest manufacturers in Sweden, right? Worldwide operations, data centers all over the world, complex sort of setup with different multiple vendors providing infrastructure and applications. And AIOps was, was identified as one of the key drivers going forward or enablers to actually make their business, which is actually we work with the IT service provider at the company, right? The business of IT simpler, more intelligent, faster, right? And last but not least, also it's, it's sort of cheaper going forward, right? So we talked about a number of things, a number of capabilities earlier as well, but to give you some examples, this new set of software, if I put it so, will provide, uh, review, uh, capabilities for new recommendations, right? So not only the software will actually be able to capture all the data, but it will put analytics on top, utilize analytics, and provide with deep insights and, and recommendations, right? Not only insights but also recommendations. So we had an issue here, uh, it actually was a, a pattern, right? A recurring pattern. So what we could suggest here is the following: either you could go press any key, right, to continue, or you should provide a more sort of CPU or disk or whatever resources to the system, right? Or you should contact your support anywhere in the world, right? So it would actually provide the operators with insights, enough insights to be able to make a smart, good, intelligent decision in timely manner, right? One of the other areas was to analyze costs, right? It would actually be able to look around and determine different types of costs, right, related to either hardware, software, services, whatever, and provide recommendations as to how to optimize those. Holy crap.
The nerd in me is like giggly excited. So what you're saying, this is like the fulfillment of a dream. It really is.
Sounds like it, at least.
I mean, no, I mean, this is nirvana. It really is. And again, is this nirvana for a person running a cleaners or a hair salon or whatever? Because in the old world, you had your IT people monitoring your network. You hoped that they would come back to you with feedback about how the network is doing. And some predict to be saying, well, we should scale the network this way, that way. We should take this system down. We should upgrade this, buy a license, right? They're monitoring and they were trying to predict it and things like that. But you're saying that that actually has been taken— is being done by AIOps. Right, and also is evaluating the network and saying that if we did this to the network or to the infrastructure, not just the network, let's be more specific, the infrastructure, that we could save money. That's like, I wish I had some sound effect, like through a party, like with sirens. I'm serious. You know, because this, everybody can get world-class service, forget about world-class services. This is the dream scenario from every, every CTO in the world has ever wanted. You can get that with AIOps. One more question to clarify something you said. You said multiple vendors before, and I think I'm probably not the only one listening that was the question that when you said multiple vendors, if I'm running my cloud or my, my storage on multiple clouds, is AIOps also able to monitor other clouds, specifically Microsoft, Google, Amazon, Alibaba, etc.? Yes or no?
It is in the AIOps console or AIOps DNA to do that, right? So that's one of the core building blocks when we actually went away and constructed this initially, right? To be able to have full view of multiple data centers, in multiple vendors, let's say, control domains, right? And of course, in a, in a centralized and unified way, right, through one single console. So that is exactly as you put it. It doesn't matter if you run your infrastructure on Microsoft infrastructure or IBM or elsewhere, Google, AWS, or your own data center for that matter, right? It has the capability to easily integrate, look into those, and it uses APIs, open APIs, so you can easily integrate those environments and have visibility across.
Oh, wait a minute. I found the sound effect. Here we go. I tell you, I'm a nerd. Oh, don't worry.
So Thomas, let me add a couple of other things which will probably make the sound effect more justified, right? Go for it. Not only the system would provide with the analysis of the costs and recommendations, but it also provides with the analysis around assets, right? Because if you are sitting with a multiple vendor setup in a hybrid multi-cloud environment, right?
Yep.
And you have various different assets which are either owned by you or you are leasing through an OpEx model or whatever. Then you absolutely need to have full control and notion of your assets. This is one thing that has been very much of a discussion point. We talked about earlier, we used to talk about CMDBs of the world. We still do that, of course, but it's very hard to find companies who have up-to-date CMDBs in-house which can easily be updated or can automatically be updated. So this system, however, will help you analyze your assets, right? And of course, in addition to that, it definitely, based on your decisions, it will also go ahead and adjust the policies across these different environments for you. So if you have different, let's say, security policies, compliance policies, all of that stuff, it's capable, fully capable to go out and make those adjustments for you, right?
And Mike, just clarify for everybody, what is a CMDB?
Yes, Central Management Database. Management database, right? So if you would have your assets in a centralized place where you actually could go ahead and do a search to find your assets and update the status of these assets, right? If you own them yourself or if you're leasing them, or if the lifetime or lifecycle of these assets have been passed or where you are in that cycle, that's the notion of the CMDB. In-house rate.
And I'm just in awe. So AIOps, I love what you said about the policy because every different contract, you have a different policy and also different countries have different policies, rules and regulations, and AIOps is monitoring that as well. Is that correct?
Yeah, it does. It does actually. What it provides with in general is a set of different capabilities, right? Such as the managing or adjusting policies, analyzing assets, you know, analyzing the costs for you, providing recommendations based on all of these, right? It also reads your problem change and all the other sorts of ticket-related issues that you have in-house, right, in your ITSM systems, various ITSM systems, right? And it also will be able to approve your orders in the end, right? So what if you need more capacity or capability in-house, right? What if you need to procure more, you know, basically computing power or storage or whatever, anywhere? It will also be able to help you doing that automatically.
Will it execute that? Will it actually execute that order and find more computing power?
If you get the workflow set up, if you instruct the system to do it for you based on a set of instructions, Of course, it will.
You know, your 'of course' is my 'wow.' Wow. I mean, really. So, if I'm starting up a company today, starting up my AIOps would be amazing. That's just one scenario. Another one is existing companies today, post-COVID, right? We're all post-COVID. The rest of our lives will be post-COVID. What am I saying? But we're looking at our existing company differently. We're looking at cost savings, we're looking at optimization, efficiency improvements. Everything you've said about AIOps just checks all those boxes.
Yeah. And that sort of notion of the idea, AIOps as such is one of the capabilities. Let me just shed some light over these different acronyms here and there and a little bit of You know, because there are different, uh, wording for, for the same sort of capability at some times. So what, what we usually use to say is that we're talking about a multi-cloud management platform which includes different sorts of capabilities, AIOps being one of these, right? AIOps itself will rely on all those other capabilities to actually provide you with all those insights and do all those sort of automated actions for you and so forth. But you still need to have that platform, which is the enabler to it, right? And having talked about the platform, you also still need to have the access to the data or different datasets, right? So those are the core building blocks of these solutions.
Right. And the platform being a cloud platform, for example.
Absolutely. Multi-cloud platform. Yeah.
Yep. Yep. Okay. Yep. Um, I had a question before, and look at it from, um, an international perspective. Uh, how, well, how do you know if the AIOps, even though it was recently announced, is it being deployed internationally? Are there many companies using at the moment?
Yeah, best of, based on best of my personal knowledge at least, there are multiple international companies adopting it at the moment or installing it, rolling it out across different platforms and environments, right? And usually the approach they take is to actually start to include their own data centers, whatever they have in-house, on-premise, as the first set out to be managed by AIOps, right? And then they extend it to different cloud environments, right? Public cloud environments such as the one IBM is offering or Microsoft is offering or AWS and so forth. So that is the approach they are taking. And some of them are pretty far down the line.
Is it? Now, here's the $2 million question. $2 billion. Let's make it even bigger. And that is, I guess I've been overexcited about this. I love technology, especially when it's so good that you can explain it in a simple way. The dashboard, because this is software, obviously, there's a dashboard that I am utilizing, I'm experiencing it through. Do I, the CTO, even the CEO who maybe only has an MBA but says like, look, I'm tech adverse, right? Everything I touch breaks. Can they— is it so easy to use the dashboard to AIOps that even a person just with a business background could utilize it and control the network, a multi-cloud network with different rules and regulations and policies and everything? Are we at that level or we will soon be there?
Yeah. So as everything else, this is software you're talking about, right? So software is an ever-evolving set of functionalities as we know it, right? So at the moment, the way this is constructed and the way companies are using it is a sort of role-based access that we can avail different capabilities in this, in the end-user dashboard as such, right? So as a CIO, you have a role defined, and based on that role, you will get access to a number of functionalities, certain functionalities. Which will help ease up your, your kind of pain points or address your daily requirements. As such, as a DevOps person, a developer, right, you will get another set of functionalities assigned to you. So when you sign into the system, you will have access to those functionalities which are relevant to your role, right, and so forth. So that's the way it is evolving at the moment, at least. And as you very well said, different functionalities in the company, different functions in the company have requirement for different sorts of types of functionalities, right? So that's the way we distinguish between these roles today, at least.
Thank you. So, but that's still great news, right? That just means that, again, I'm coming from the startup side and I know a lot of people listening are running big companies. The leaders who are listening to this in the big companies, That means that you can actually get more from your team. Just a smaller group of your team could actually run your entire network. The rest of your team can actually take your company to the next level. This is great for them as well as a startup where maybe just you having your CTO and one or two other nerds is good enough to scale to whatever size you need to scale to. I actually see no downside to any of this. Mike, I think we got people interested. And also curious. How easy— if somebody out there has no relationship with IBM at this very second, but they want to at least have a discussion about AIOps, or a lot of people just say, "Let me try it before I even talk to an IBM person." That's fine. Where do they go? How do they go about doing this right now?
Right. So there are a couple of easy ways to do that, right? Either you can contact IBM Cloud and Cognitive entity. And then you'd be talking about— so, so this is— let me, let me just pause a little bit or, or, uh, rewind a little bit and, and mention that this comes, uh, AIOps comes as a part of MCM, right? Multi-Cloud Management Capability, the platform that I was talking about earlier, right? Yeah, not as such can be delivered or is delivered in two different ways today, mainly either as a service by IBM Services, right? And that's something you can subscribe to, and then IBM will take care of all the integrations for you in the background so that you can actually subscribe to this service and just log on and have a view of your multiple workloads across different environments, right? That's one way of doing it, and that would be a contact which needs to be made to IBM Services, right, to through your client executive. The second way of consuming it is through installing the software yourself, and then we call it multi-cloud management Cloud Pak, or Cloud Pak for Multi-Cloud Management, right? And that's basically availing you to be able to install the software, set up your integrations yourself or with help of IBM through consultancy services, right? And then you own the piece of software and you can actually adopt and tweak it yourself on your own.
So let's simplify that even more. So there's two ways. If you're listening to this and you really want this, this is something that you want for your company, or at least just to play with, or you want somebody in your company to play with, there's two ways to contact IBM. And one is that is to contact your existing contact at IBM if you have one, more in IBM Services, if you know anybody there. If not, just contact IBM and they'll put you in contact.. But if you have an existing IT team and they say, you know what, just we just want to play, we won't talk to anybody just yet, then the— which cloud pack was it again, Mike, to be specific?
Multi-Cloud Management.
The Multi-Cloud Management cloud pack. Fantastic. So there it is. So either you want to work with IBM directly, contact somebody, or you want to play with on your own and then engage with IBM. Those are two different ways. And both ways you'll have access to our latest and greatest announcement that I'm really trying to I'm flaws with, but I can't come up with any. Mike, let's be honest. Let's be transparent because nothing's perfect. Are there any flaws where AIOps— and the flaws could be this— is brand new announcement. What are the new things that are coming up? So that's not a flaw, but what do you know with the next version, AIOps or whatever, what's coming down the road to get people more excited?
Right. So very good point, Thomas. And let's be very open about it, right? Nothing is perfect, and this is an evolving piece of software we're talking about, uh, although it's, uh, it's been out for a while. Uh, we— let's talk about a little bit about integrations, right? Which APIs do we have in-house today? So we can, for instance, today, uh, provide with APIs to connect to ITSM tools, toolkits such as IBM Control Desk, ServiceNow, VMC Remedy, and a number of others. But are all integrations in place or all APIs in place today coming out of the box? No. So this is a point where we are sort of working very fast to develop new APIs and set up integration points, right? So that's one of those points. The other one is, do we have all dashboards set up for all different types of capabilities that I was touching upon earlier? Are they all developed? Probably not, but we are working on it. So as mentioned, everything should be in place for the portion which is in-house, meaning classic data center and cloudified environments in, in own data centers. But as for different cloud environments, and particularly when we talk about the different container environments and so forth, this is something which is under development. Rapidly developing.
I think that's great. I mean, every— look, this is cutting edge. This is just right out of the box, right? So that's absolutely fine. There are growing pains, but all the things that this helps a company is more than worth it. Mike, I have one more question before we go. And I was thinking about Red Hat. IBM acquired Red Hat, as most people know. Is there any synergies or integration synergies with AIOps and Red Hat?
Yes, Thomas, uh, we talk about containerized environments, right? So as Red Hat is adopting and building on open technologies, right, and as AIOps is also working with open technologies, there are many synergies in between. I, I mean, I think in respect of time, we don't have time to deep dive into that, but, um, definitely so, definitely so.
The reason why I bring that up is I thought about it spontaneously, that's actually a third way in which people can engage with AIOps. And that is because Red Hat is now owned by IBM and you have an existing relationship with Red Hat, you could definitely expect, if not already, that AIOps and Red Hat will be integrated or compatible or whatever. So it's just, and Red Hat has such a large footprint worldwide. So I'm trying to get people more excited and give them less reasons not to play with AIOps. Because, you know, the power of IBM is everywhere, is doing a lot of amazing things. And the move with Red Hat was genius, right, to bring services like AIOps to everyone without them having to make any huge changes. So Mike, you get the last word. What is, what is the last thing, or do whatever you want, say whatever you want, just don't be insulting to me.
Thank you, Thomas. Uh, I think we've, uh, we've talked about the most significant parts of AIOps and multi-cloud management platform, which AIOps is a part of indeed. Uh, and, uh, what I would like to say is basically the notion of classic data center management and IT management or IT service management is, uh, basically disappearing as we speak today, right? Uh, with that said, I'm not saying it's disappearing for good, but it's going to be part of a much larger thing, which would be multi-cloud management going forward. And doing multi-cloud management, we better do it the right way, meaning we better infuse, make sure that we infuse cognitive capabilities from day zero, from the inception phase, so that the day when we stand there, the day it's such a complex and multi-vendor sort of related environment, then we are standing ready to be able to not only manage but also take the most best advantages out of it. Right.
Thank you very, very much, Mike. I just want to emphasize to everybody, listen, this is Mike. It's been great having you again. Mike Gorashi, say his name the right way, everybody. IBM CTO Nordic of Cloud Acceleration and Transformation. Mike, thanks so much. To all you listening out there, From all of us at IBM, please take care of yourself. Please be safe. That's on a personal level. And now from a business level, listen, times are hard out there. We get that. But, you know, we brought Mike onto this podcast right now at this specific time because what he's been working on with AIOps is really something you've been looking for, something you need to simplify your life, simplify your business. So, Mike, thank you so much. I appreciate it. I look forward to talking on internal calls and laughing some more. To all of you out there, take care, be well. Mike, say goodbye to everybody.
Thank you very much, Thomas. Thank you.
Take care, everybody. To the next time.
Bye-bye. Cheers. Bye-bye.
More from Thomas

IBM Watson AI Lab for Cancer presentation to Ullern HIgh School / Thomas Anglero

Will You Lose Your Job to AI / Oslo Business Forum / Thomas Anglero & Richard Qwest








