Austin Pierson | ScienceLogic | IP Expo

Avatar Ben Fower | 13/11/2018

Austin Pierson form ScienceLogic joined us at IP Expo to talk about AI Operations and how they can benefit any company. Hosted by Allan Behrens and Matt Lovell

Matt: So, Austin from ScienceLogic, just help the audience understand a bit about yourself and your role in Science Logic.

Austin: Sure so, I look after ScienceLogics business in the UK and South Africa more from a commercial perspective, although I’ve got a bit of a technical head and we’re one of the leaders in AI Ops. In fact, AMA radar came out with a new report this week puts us in the top three vendors. So exciting space to be in yes.

Allan: Explain that, expands that if you could.

Austin: So, a couple of years ago Garner decided they would take four or five different bits of monitoring essentially. So the basics of monitoring and event management and machine learning that kind of thing, put it all together in one in one group and call it AI Ops, and so essentially it’s the automation end of monitoring so making sure that everything is available wherever it is in the cloud in your datacenter, whether it’s an IOT device or traditional IT and making sure that not only is it available but you’re taking actions to automate things when things go wrong as they inevitably do and then using the magic of things like machine learning and AI to predict the future when you’re going to have a problem and fixing it before it before it happens.

Allan: I like the concept of fixing it before it happens. They used to have a term called autonomic which is fixing it after it happened.

Austin: Yep. Absolutely.

Allan: Right, We should think about it as a path later.

Austin: Yeah. Absolutely. Yeah, the future is pretty interesting. So, we got a large hook up with IBM and their Watson platform at the moment. So, taking in not just IT data, but external data so how does weather patterns predict IT usage? So, if you’re a retailer, for example, you’re going to get more traffic on a rainy day. Therefore, your IT systems potentially need more resource. So, you need to chuck stuff at it before the event happens and things like that. So, it’s a really interesting kind of use of technology, I think anyway.

Matt: Okay so the geography is an interesting part. Technology is evolving so quickly obviously organisations had more SasS platforms and services or interfaces greater coalition of data. How does AI odds keep up with all the change?

Austin: Yeah, so great questions. So, I think there’s been a massive, massive shift in this kind of technology. If you went back probably four or five years ago, you know, you were talking to four big vendors. They’ve been doing stuff for donkey’s years and as soon as something new came into your environment, you waited 12 months for them to develop a new connection to that agent it probably was in those days. So nowadays, thankfully certainly our software is very API driven and that’s the new world. So great you’re taking on something like Google Cloud. All we do is just hook into that API straight away and within minutes you’ve got that connected into your ecosystem. You’ve got it in your business service mapping topology all automatically. So, it’s really really simple to put that technology and to integrate it into your environment easily.

Allan: I mean, do you provide the service to learn the environment or is that something your customers has to do?

Austin: No, so that’s all automatic. So that’s one of the great things is on day one we would discover automatically all of your environment wherever it is, wherever your technologies are and how they’re talking to each other and then we keep that updated on a less than five-minute basis. So for things like containers, you know kubernetes, v-motion and whatever in the past that was just like a guessing game now, we can tell you straight away and I think one of the big growth areas we’ve seen with customers is then keeping that up to date in something like ServiceNow so that your CMDB is up to date in real time, not in a weekly or overnight scan because that’s no use anymore. So, therefore, when an incident gets created it’s automatically got the asset in it. It’s got all the data that you would need about that business service what service users are affected. And you can automatically push that to the right person to fix that thing very very quickly. So, our biggest customer is Cisco. So, they use our technology both for their internal IT and for their managed service and they use our stuff to collapse 53 different monitoring platforms into 1 so they use us with ServiceNow and IBM Watson. And their mean time to resolve was two hours and 20 minutes before they put our stuff in. Complex environment, right and they’ve got a lot of Telco they’re working with and things now they’ve got it down to six minutes.

Allan: Wow, that’s quite a benefit.

Austin: It’s not bad, it’s not bad.

Matt: Six minutes through its learning and understanding, you know behaviours, events and instance of the problems or is that you actually taking a process breaking that down and saying okay time to resolution all the steps and automating it.

Austin: Yeah. So, the first thing we do with any customer no matter how small or big is we go through that incident or event to incident process with them and say, okay, where are the problems? Where are the quick wins going to be? And number one is what’s the first 20 minutes of any service desk call is what the hell have I got and who’s affected by this? you know, sort of thing so automating those pieces is clearly very cool. But what we’ve done with Cisco is being able to take that a lot further to the stage where we’re going. Okay, what questions do we ask a device when we know there’s a problem. Ah right, we’ll go and ask you this, this and this, but let’s just forget that let’s just automate all of that. So, our team up with Cisco is great because any customers now we have that have got Cisco technology, 99% of the world pretty much, they can take advantage of our relationship there and take those automations and plug them in straight away. So that on day one they’re going ah right okay, we’ve got a problem with the Cisco device, ah we already know exactly what the problem is. We’ve asked all these questions great and we know how to fix it automatically.

Allan: I suppose part of that is down to the vendors making their devices smart enough to furnish the information.

Austin: It is.

Allan: So what advice would you give to vendors bringing out equipment?

Austin: Work with people, with other vendors such as ourselves before you’re bringing the stuff out. You know, I was over at the Cisco Innovation stand over there and they were talking about some very cool stuff they’re doing in the future and we’re getting involved with that but there is no reason why you can’t give pre-release betas to vendors such as ScienceLogic and say, hey look this is what we’re developing, you know, we’re part of your ecosystem. We’re doing that with Nutanix of the moment. Obviously, we’re doing it with IBM and Watson and things and it makes such a difference because then when they bring out the next version of that technology, they’ve already got an ecosystem of partners that support that which means their stuff will fly off the shelves a hell of a lot easier.

Allan: Are there any standards that are being developed to support people who perhaps don’t agree with betas because obviously, some people are quite…

Austin: You know what pretty much, the API world is changed that kind of, I guess, ecosystem family development so much as long as you’re building yourself with an open API, then realistically it’s not going to take more than 48 hours to plug into a platform like ours where everything is driven by open APIs.

Allan: Right, right. I mean just out of interest. You know, I know that this open API is a big thing. Do you still see a resistance by some companies to change to open APIs?

Austin: No.

Allan: Good. Well, that’s good.

Austin: No, Not really.

Allan: If I’d have asked that question two years ago?

Austin: It would have been very different, but I guess you still get a pushback in some environment, so talking to a big bank, for example, you know, they’ll have plenty of their old ways of doing it. But even those guys they’ve got innovation funds and department and those guys are going actually ‘we have to do this’ you know, I’m working with some guys at Manchester at the moment, you know, and they’re just setting up new divisions to go okay, we can’t work in the old way. Let’s set up a Dev Ops department and that’s where all our new projects go through and they’re like, yeah, of course, we use APIs. Why wouldn’t we?

Matt: Interesting point of view because so much of organisation or effort is now going into infrastructure as code and building in quality assurance, quality testing into the codes. You’re monitoring the API and the interrelationships between services whether it’s defined network location services, the infrastructure services, but how do you combine that with how organisations are now either writing services themselves micro or otherwise or they’re deploying infrastructure?

Austin: Yeah, so I think interestingly. It’s actually made the landscape more complex than it used to be because it’s so easy to do this stuff now and there’s so much cool stuff out there. You write something, chuck in a container or whatever and then you’ll connect it to AWSs amazing video services that are then recognising people are walking past your shop and saying yep they bought from us before let’s send them a text sort of things. I think more need for things like ScienceLogic than there used to be when you had a physical server and another physical server which had a database on and then your web server sort of thing, actually life was dead easy then now you’ve suddenly got this massive ecosystem of services that you’re plugging together and if one piece of that ecosystem doesn’t work. You’re in complete trouble so that’s where we’re seeing a real real growth area for us.

Matt: How’s that challenging the way you present the information? Because you’ve got this transition to service monitoring. You’ve got all these artificial intelligences, a huge number I guess of different inputs and end points that you see in that process, you know, occur over and over and it’s learning from at the same time. How do you go about presenting that as a simplified deal?

Austin: Yeah, so very much five years ago we would have sat there and sat here and talked about business service management. I think it would have been the view that everyone went to people who invested a lot of time and money into it and then went this is way too hard, you know, we’re never going to get anywhere. Actually, it’s pretty simple to do. So, that’s where we would start with customers, forget the old stuff. You know, what is your business care about and let’s put that on a dashboard and go here’s retail stores, here’s online commerce, here’s how your Warehouse is working, here’s your payroll. That’s the exciting bit, soon as you click on that underneath then you get to the next screen of ah right some of its in AWS over here some bits are easier and there’s VMs over here or whatever, you know, but it’s layers.


  • © 2019 COMPARE THE CLOUD LTD. All rights reserved.
X
X