Episode 30 of The Andy Show, Andy is joined by Kevin McGrath from Spot
On episode 30 of The Andy Show, Andy is joined by Kevin McGrath, Chief Technology officer at Spot..mp4
ANDREW CLEAN [00:00:45] Good afternoon and welcome to The Andy Show. The day is the 1st of June, 1st of June. I cannot believe it is our 30th episode. It is, time is coming up to 12:31 in the afternoon in the U.K. I say U.K. because my guest is coming to me all the way from the other side of the ocean. It is my absolute pleasure to be joined today all the way from Maryland by Kevin McGrath, the Chief Technology Officer at Spot. Kevin, welcome.
KEVIN MCGRATH [00:01:20] Andy. Thanks for having me.
ANDREW MCLEAN [00:01:23] No problem at total. No problem. Oh, I seem to have dismounted my microphone. So, Kevin, I’ve been reading about you I know that you’ve got a very colourful past in everything from radio to technology. So would you like to start by telling us a little bit about yourself and what you do at Spot.
KEVIN MCGRATH [00:01:44] So, I’m the CTO at Spot. I come from a long line of way back in the day software as a service before or applications as a service before we were calling it software as a service and at Spot, what we do is we really look to optimise and make people’s Cloud infrastructure as efficient as possible and run production-grade applications on top of excess compute capacity in the clouds.
ANDREW MCLEAN [00:02:09] Fantastic. So let’s start at the beginning. Spot, I am aware of Spot and I’m sure you’ve changed name at some point. You were, it was Spotinst?
KEVIN MCGRATH [00:02:21] Yes, yeah. So we started as a Spotinst and my mother was not able to pronounce that name. So we decided to shorten it up, much like other companies over time, which have started with longer names and they kinda shorten down to like the thing that they’re known for but we’ve shortened our names to Spot recently. It’s just a couple of months old now and we really want to be, you know, to say the one spot that you go for all your optimisation needs and the Cloud. So we’re rolling with that. It’s a much easier name to say going forward.
ANDREW MCLEAN [00:02:52] It’s a great name. It’s a great name, spot.io is the website. So, Kevin, we were chatting earlier and I wanted to say to you that, I mean Cloud now more than ever has become so integral to everything that’s happening, particularly during this lockdown. Suddenly local servers, have become a little bit difficult, especially if they were in people’s offices and there’s been a huge rise in Cloud and the amount of usage that’s been going on. Before we talk about the Cloud journey and how things like lockdown have affected that, do you want to tell me a little bit about the background of Spot because I believe it’s quite an interesting story, the kind of founding points.
KEVIN MCGRATH [00:03:34] Yeah. So Spot, itself has a really fun journey back. Well, back in the day, our founder Amiram, worked or was in the Israeli military running data center structure services for things that support services they have like the Iron Dome. They were looking into Cloud services at the time. So he started, you know, kind of that hybrid model very early on in his career and some of the people that he was working with all went to do their post studies afterwards and while he was in school, wrote a paper on how to run production grade applications in the Cloud. What he realised is when they were moving these applications from the datacenter to the Cloud, it wasn’t necessarily costing any less without a full re-architecture. While you can gain all the benefits of, you know, software during development and using things as API in the Cloud, some applications don’t actually cost less unless you really kind of rethink the model when you get to the Cloud. However, using excess capacity by the Cloud, which are called Spot instances in AWS or preemptable instances in Google GCP. If you can figure out how to use these instances which the Cloud providers sell at fractions of a dollar to you, then you can save up to 90% off your bill. The problem is the Cloud providers can take these instances back from you at any time. So what you need is you need some sort of model, some sort of learning that will know the capacity of these markets and when, these instances are gonna be taken away and then you can keep applications running at, you know, near full ability, near 100% reliability. You can put a 49 SLA on these applications and you can keep them running as the infrastructure changes underneath. So he wrote a paper on this, very early on after getting out of the military. That paper has turned into a company and what started as really taking advantage of this excess capacity has turned into a full optimisation and efficiency story, not just for compute, but for total Cloud spent for very large enterprise organisations.
ANDREW MCLEAN [00:05:35] It’s really interesting. I like the idea that it’s been the “boffins” have taken it to the next level from the theoretical to the application. So you’re in a somewhat unique position from the people I normally talk to because you’re in the US, so it would be really interesting to get a flavour of the US and Cloud adoption. So, I suppose there’s a couple of questions there. Has… I believe there’s been quite strong Cloud adoption over the past maybe, decade. But also what kind of companies have you been working with in Spot?
KEVIN MCGRATH [00:06:10] Yeah. So that’s one of the more interesting things. I believe very early on you were looking at companies who were just looking to grab as much compute as possible and didn’t want to outlay the CapEx. So very early on, you’re looking at a lot of ad tech companies and companies that need to scale very quickly on demand. But I think that’s the type of persona you are looking at maybe 4 or 5 years ago. Now you’re seeing almost everybody and it’s not everybody’s giving up they’re on prem data centres and moving completely to the Cloud but almost everybody has some sort of Cloud piece to their puzzle, especially large enterprises. You see a lot more adoption going to the Cloud because it’s quicker to spin up. You get that, kind of, all these API, you get developers can get started very quickly. Without, you know, without a lot of wrangling internally and setting up, you know, full teams to stand up, you know, physical infrastructure in the API’s on top of that infrastructure and having a whole team to support that. You can click a button and have all of that available to your teams right away and if you think about how to run that cost-effectively, you can run anything in your, you know, anything your data center you can move to the Cloud and now that you can, you know, hook layer two up between, layer two networks up between your data centre and the Cloud, it feels like, you know, the Cloud is just an extension of your home base where it used to be. So now you see retailers, you see ad tech, you see financial companies, even people with data sovereignty are bursting to the Cloud to run compute. So even if their data sovereignty has to be on prem, they’re using the exccess compute in the Cloud. So it’s really all walks of life moving to the Cloud at this point time.
ANDREW MCLEAN [00:07:51] So when I look through your site the other day, I noticed some very big names on there. One that struck me straight away and I’m asking this because of the amount of demand that they had, well maybe a little less at the moment, but generally the amount of demand that they have, which was Ticketmaster, one of your case references. Can tell us a little bit about that and why Ticketmaster got in touch.
KEVIN MCGRATH [00:08:16] Yeah, Ticketmaster is a great story for us. They use a product called Ocean and they run their Kubernetes, they run a lot of Kubernetes and as you can imagine, they have a lot of spikes in demand when ticket sales get really high and they have to get, you know, lots of compute at different points in time. So, when they scale in the Cloud, it is Spot. It is Ocean by Spot that is scaling those infrastructures when ticket demand goes up. When they need more compute capacity. That is Spot in the background. That’s finding the extra capacity at the best price and best availability for them and as they look at it, they can lean on us, provide all of their infrastructure while they focus on delivering their products, right? So they’re not worrying about how to get infrastructure, how to get instances, how to get them at the right price at the right time, how to make them highly available. They put all of that on us. We’re making sure that the infrastructure is available at their SLAs and SLOs and they focus on their applications. So they focus on writing code to sell tickets. We focus on making sure that code can operate.
ANDREW MCLEAN [00:09:25] It’s really interesting because a lot of companies that founded in the Cloud, they had a similar message to begin with about reducing cost, reducing costs. Then I noticed when you were talking about when you initially started up, one of the things was to bring the cost down. But you’ve kind of it sounds as if with Ticketmaster, for example, you’re kind of moved into all sorts of different areas of why people are using Spot. It’s not just about cost, is it?
KEVIN MCGRATH [00:09:52] No, it’s a great observation. As people move to the Cloud, you’ll realise that you eventually start redesign your application so they’re a little bit less tied to the infrastructure than they were before. So these applications can naturally survive kind of instances moving in and out or auto scaling and things like those things that maybe 10 years ago not every application was programmed to be an auto scaling application. But nowadays, especially with “containers” and Kubernetes and where we see the market going with that, more and more applications are quote unquote “Cloud ready or Cloud native” at this point in time. Now, that’s great for the applications, but the infrastructure still needs to be there and it still needs to be available when you need it. So we’ve really found ourselves going from just going to find a cheap instance for you to run your applications on, to keeping, you know, large scale applications alive and running and integrating into these different tools that customers use. So integrating directly into Kubernetes, integrating directly into things like “doctor swarm” or other scaling applications or into your CICD flow, into your Jenkins’ flow and things like that so you can run like really large scale tests that take thousands of servers. We integrate into all of these points of contact with our customers. So it’s very just natural for them to use. They use these applications and not caring about how they provision the infrastructure on the back side. We take care of that.
ANDREW MCLEAN [00:11:19] Beautiful. Well, I want to talk a little bit more about some of the things that Spot are doing now, particularly interested, you mentioned it there in your Ocean services, which ties nicely into Red Hat that we’ve spoken to recently, actually. So I want to come back to that. At the moment, though I want to move back into the generalised topic around Cloud in the current environment. So we’ve, all of a European in America, we’ve been hit with things like lockdown. We’ve had the frantic rush for people to work from home. The changes that come with that and possibly a future where working from home will become a bit more of the norm. How have Spot seen a change in demand since this whole thing started off or has it just been normal?
KEVIN MCGRATH [00:12:16] There is a change, there’s change for everybody. Let’s not kid ourselves. I get on the phone with people from all over the world now, whether it’s South America or Europe or the United States. Everybody is working from home. It’s all these home backgrounds that you see now. It’s, you know, not everybody’s in the office. Things have changed. Now for us, for Cloud, Cloud has seen, I would say, an increased adoption. We’ve had customers at all ends of the spectrum, customers who are obviously maybe in, you know, brick and mortar retail that are hurt a little bit more than people who are on online, retail and growing. Generally, though, we have seen the use of Cloud increase and the one thing that we have seen, though, from companies is the current volatility and expected volatility of what’s going forward and the unknown and because of that unknown companies are being a little bit more restrictive of their pocketbooks and their spend at this point in time. You’re going to see a lot of less probably large capital expenditures, which is probably why you’re seeing a much larger uptick in Cloud. Because the Opex cost is something that they can shell out now and, you know, pull back in a few months if they need to or continue to go, right? Because it gives them a lot more flexibility and because of the volatility of not knowing where people are and where they’re going, especially even growth companies, even companies that have seen a huge increase due to COVID are very concerned about where their money’s going and how it’s being spent, especially over the past two months and the next six months going forward and we’ve, you know, we’re in a lot of opportunities right now where we’re helping people control these growing costs due to the volatility that they see. So beforehand, where you might have just had kind of like a blank cheque to go out and use the Cloud and run the application you wanted to, these budgets are being pulled back and forcing different groups to use the Cloud to make sure they’re using it in the most efficient way.
ANDREW MCLEAN [00:14:08] I’ve been hearing different stories from people regarding projects, a lot of people have very big projects. And I would have thought when they spoke to me, many people spoke to me about this, but they would’ve said, well, all, the projects are on hold. But actually, the availability of time, which is obviously shifted somewhat, has meant that there are a lot of projects up suddenly just gone boop. Have you seen that over there? Have companies when you would have assumed that they would have, right, we’re just gonna keep running and just ramp up for this. But actually, they’ve started new projects.
KEVIN MCGRATH [00:14:43] Oh, yeah. I mean, I think everybody is getting creative in the different ways that they look at things, right? Like we had, we definitely had some customers who were hit fairly hard early on who have kind of come out of things now and are back on the uptick and, you know, we’ve been there with them the whole way, you know, either, you know, helping them out any way we can, realising that, you know, COVID is a situation that we all have to deal with. So we work very, very closely with some customers who have shown a dip. We haven’t had you know, I knock on wood and everything like, everyone who was our customer before COVID is a customer, you know, currently today. We haven’t, fortunately, we haven’t had someone go out of business or had to cut ties or anything like that. We’ve been able to keep a pretty good story going with our entire customer base. But people have had to get creative, right? Applications that they weren’t necessarily running with our platform today they started running with our platform to try and save money. They had to do some changes in how they do things internally. They’ve had to decide maybe not to do some projects and do other projects. It’s been big. You have to take a look at what the market’s giving you now and then you have to look into the future and realise COIVD is going to be with us in some form of capacity for the next 8 to 12 months and you need to plan your business accordingly and that’s just the way it is, whatever your personal views are not on whether people should be locked down or if things should be open or how you want to deal with that. The reality of the situation is the governments from all over the world have shown that this is going to be a thing and this is gonna be a thing until a vaccine shows up. So you’ve got to deal appropriately and every business is different. There’s no one answer. I can’t give you a single answer, but everybody’s plans have changed.
ANDREW MCLEAN [00:16:22] And the question I ask a lot of people that come on the show. What about Spot themselves? How have you guys coped with the lockdown, with the changes, with the remote working as an organisation?
KEVIN MCGRATH [00:16:35] That’s a great question and in the United States, we are a little bit more accustomed to kind of the remote work. I myself was I’ve been on and off remote worker for a lot of times I travel. The one thing that I don’t do anymore is that on travel to all the offices and getting face time with people. But we have transitioned out of the offices and we’ve gone completely work from home. I will say that the employees here at Spot have been awesome about the entire thing. We weren’t traditionally a work from home company, we had a lot of people, especially our dev groups in like, Israel, would come to the office as a very natural thing for them to do. But we’ve been working from home now for at least a couple months and we’ve kind of had to band together and change some of our own internal communication policies and how we do things. But we have a great leadership team here and we have even better employees and everybody stepped up. Everybody knows it’s kind of a different situation. And we do a lot of things remotely now. We do a lot more writing. We do a lot more, you know, planning out what you’re going to say, communicating with people differently. We’ve been really, really stringent on making sure that forms of communication are written down and you can comment on them and large groups of people can see them. And we use a lot of, you know, video conferencing and we try and get in face time where we can and still have fun, right? That’s the other thing. Getting the groups of people at the company together, having happy hours, having toasts, celebrating people’s birthdays and things like that is also very important, something that, you know, you could very naturally do in the office just by tapping a bunch of people on the shoulder. But we have to plan a little bit further out just to kind of keep the community going. But I have to say, we’ve been doing great. We’re still growing, right? Our revenue is still going up and our headcount, we’re still hiring through this entire thing. So, you know, anybody out there is looking for a job. Spot is hiring all positions, sales through engineering. So come on out to our website and send over your resume.
ANDREW MCLEAN [00:18:29] There you go. Get in touch with Kevin, get in touch with Kevin. So, yeah. Going back to the technology, one of the things I’ve noticed even myself is that automation and monitoring have become more important now than ever. I think previously had been I just rely on someone to do this and this but I think intelligent monitoring and a certain level of automation has become more important during this period. Is that what you found?
KEVIN MCGRATH [00:18:57] Yeah, absolutely. I think that’s a that’s a general story for Cloud, Cloud in general, right? You have, one of the things when you move to Cloud, you have all of these different pieces that start fitting together like you might have, now you have a Cloud provider that you work with, not only Cloud provider, but you might be tying together 10 to 15 different services along with your application that currently runs and you need to monitor all of these things. There has to be someone who can kind of sit in the middle and say, like, hey, are things healthy? What does healthy actually mean? You got to call in metrics, monitoring tools. Your application relies on a lot of different services that are running once you get to the Cloud. So it is important to take all this in and then you need to use those metrics. Am I going to decide to scale? Am I going to side the scale up or scale down? Do I need more infrastructure? Or what’s my runway like for the next you know, the next few months on, you know, how all these different inputs affect my application. And then on top of that, you add in things like Kubernetes, where you get down to a container level, which is, while it’s great and it’s become kind of de facto way for us to deploy applications and run things at scale, it’s added a whole nother level of metrics on what’s my pod doing? How’s my Kubernetes cluster interfacing with the rest of the services that I’m running in the Cloud? And how did the services and Kubernetes work with the services and my Cloud? And how do I scale on that? How do I tie my storage together or my security together? So what we haven’t done is we haven’t reduced the amount of endpoints that we have to monitor in the amount of metrics that we have to look at. So it is very important to pull all those together. And it’s something that we do when we look at, you know, how do we scale something like a Kubernetes cluster? You have to take all the Kubernetes metrics into place plus your Cloud provider metrics into place and then make intelligent decisions based on those metrics coming in.
ANDREW MCLEAN [00:20:43] It’s a… One of the thoughtful discussions that always come around Cloud, I can’t keep up. I cannot keep up with this. People talk about private Cloud, public Cloud, hybrid Cloud, blah, blah, blah, blah, blah and everyone I speak to it has a kind of a different view of it. I mean, is public Cloud still going strong?
KEVIN MCGRATH [00:21:07] Yeah, I would say public Cloud- I would say all I.T. in general is very strong. All right? So the entire I.T. industry is on an uptick. I think you see increases. I think public Cloud is still on a faster growth plane than on prem, but on prem’s on an uptick too. I think on prem spending has gone up over the past 5 years, even with you know public Cloud adoption. So what you’re looking at is I.T. in general is still on a huge growth pattern. It’s still on the hockey stick curve. Not all companies have, like, fully bought into, like revamping their I.T. structures, as you see that I think you’re going to see public Cloud grow at a faster rate than on prem. But you’re gonna see both grow and what you gona see, is you see companies intelligently decide where to put their applications. Some people on prem still going to make sense. For a lot of people, public Cloud is going to be the de facto. You know, these born in the Cloud companies of the past 7 years, they’re going to start on public and then maybe they will sprinkle in some on prem. But you’re gonna see people pick the technologies that are best suited for their business. And that’s going to be with Cloud providers, too. You’re gonna pick, know, one public Cloud provider over another Cloud provider based on the services that run your application the best way. So I really see a growth pattern in all of these. And you’re gonna see companies needing to tie these together, maybe not necessarily run the same application in a bunch of different locations, but you’ll have multiple applications. One might run on prem, one might run of a AWS, one might run in Google, one might run in Azure, all for different reasons. And you’re gonna need a way to kind of tie the visualisation of these applications together. One thing, our platform also helps us.
ANDREW MCLEAN [00:22:51] So my producers send me a couple of things that Spot are doing. I just gonna ask you what they are and I hope I get the pronunciation right. Funny, I was watching a program recently and it is a different pronunciation than I’ve ever heard and I think that was I played American at Eco.
KEVIN MCGRATH [00:23:11] Yeah. So yeah, it is. It is Eco. Eco is a way for us to automatically manage Cloud commitment so reserved. If you’re familiar with AWS, the easiest way to explain this is reserved instances. But there’s reserved capacity on all the Clouds. Azure, Google and AWS reserve capacity is basically saying, you know, I’m going to pay you an upfront “fee” or I’m going to commit to spending this much on a certain type of compute over the next year. So the Cloud provider is going gonna give you a discount. Now, what normally happens is you’re locked into that for 1 to 3 years and you can’t change it. But what Eco does is Eco can actually buy and sell this reserve capacity, especially in AWS, on a secondary market. Which is supported by AWS. So if you come in and buy this reserve capacity, which you pay 40 cents on the dollar for, then you can feel comfortable that you don’t necessarily have to commit to it a year or 3 years. Because if we notice you’re not using that commitment anymore, we can sell it to somebody else and we can also buy from other people on the marketplace. So we can move these commitments around for you. So you can still get the savings for traditional applications that don’t scale too often, but you’re not necessarily locked in for the period of time that you’re normally locked in. So it’s it’s savings with flexibility all at the same time and it’s usable for any application.
ANDREW MCLEAN [00:24:33] Perfect, and the other one, which I’ve read about recently, tied in with the Red Hat. You’ve had a Red Had announcement and of course, that’s was acquired recently-ish by IBM, it’s your Ocean. Can you tell us what Ocean is? And can you tell us a little bit by the Red Hat announcement?
KEVIN MCGRATH [00:24:53] Yeah, absolutely. So Ocean by Spot is our Kubernetes infrastructure auto-scalar. We basically provide the serverless experience for your Kubernetes cluster, but everything still runs within your own account. So you don’t have to, like, give the keys to the castle to somebody else to run all your infrastructure. We run infrastructure within your Cloud account, but we give you that serverless experience. So you don’t have to, you know, manage the instances. You don’t have to actually manage to compute that goes into serving the containers that you’re running. And the integration that we have with Red Hat is we have an official Red Hat operator. So with Red Hat OpenShift, whether you’re running your cluster on prem and then you burst to the Cloud or you’re combining different Clouds together or however you’re using open ship today, Spot, or Ocean by Spot and sit there as a controller, you can use it through your native “Kube CTL”. “Kube Cuttle”, however, you want to pronounce it again with pronunciation things. However you use Red… OpenShift, naturally, Spot can be there making all of the intelligent decisions in the background which compute to get at the right time for the right price and the right availability. So you’re not provisioning infrastructure. Your developers, your engineers, they’re focussed on deploying applications with OpenShift and then Spot is sitting there in the Cloud provider of your choice, making sure that infrastructure is as efficient as possible.
ANDREW MCLEAN [00:26:18] And as part of your move into this for taking in multiple public Clouds because a lot of companies have the public Cloud of choice as they say Azure or AWS?
KEVIN MCGRATH [00:26:30] Yeah. So we run in all the Cloud, we run in Azure, we run in GCP, we run in AWS, we support Kubernetes clusters, all there. We also support, we support AKS. We support GKE. We support EKS in all the Clouds. So whatever, whatever your choice, whatever your flavour of Kubernetes orchestrator, you know, you’re you’re using OpenShift. For instance, we can hook into any Kubernetes at point. It it doesn’t matter where it is or how it runs. We don’t concern ourselves with running that control plane with one running the master servers. There are people who do that very well. We don’t get in the way of that. What we do is we run that data plane and that compute plane as efficiently as possible for you. So while your management playing probably 3 to 5 servers that are really the brains of the operation going on, and we’re all the scheduling happens. We’re managing that large compute data plane which can scale to thousands and thousands of servers where the applications are scaling and running.
ANDREW MCCLEAN [00:27:24] Okay. Kevin from Spot, thank you so much. You… this part of the interview is now over. You’ve talked about Spot. We’ve spoken about other things, but there’s a final bit in our final two minutes. And it is our psychometric quiz. However, there’s a problem with the psychometric quiz, it’s very British. Have you been to Britain many times?
KEVIN MCGRATH [00:27:46] No. No, I haven’t.
ANDREW MCLEAN [00:27:48] Okay. Excellent. So with the “phonology” head, which is an outdated Victorian concept, we’re going to ask you some quick fire questions. You can rationalise your answers if you want. There is no right and no wrong answer. I’d I’m thinking of course you know what I’m talking about. Okay, so I’ll start. Just two questions where everyone comes into your head first. So let’s start with an easy one. Paris or Los Angeles?
KEVIN MCGRATH [00:28:13] Los Angeles.
ANDREW MCLEAN [00:28:14] Woo! Why is that?
KEVIN MCGRATH [00:28:17] Stay with. I just you know, I’ve I have some friends out in L.A. I been there a few times, you know. I’m not going to stay U.S. on this one.
ANDREW MCLEAN [00:28:27] Okay. Oh, cheese and onion or salt and vinegar?
KEVIN MCGRATH [00:28:32] I’m not a vinegar fan, so I’m going to go cheese and onion.
ANDREW MCLEAN [00:28:35] You call them chips? You call them chips over there, do you?
KEVIN MCGRATH [00:28:38] Yeah. Yeah. Yeah, I guess. Yeah.
ANDREW MCLEAN [00:28:44] Brown sauce or tomato ketchup.
KEVIN MCGRATH [00:28:48] Oh, that’s a good one. I have four kids. I eat a ton of ketchup. I’ll go ketchup on this.
ANDREW MCLEAN [00:28:55] EastEnders or Coronation Street?
KEVIN MCGRATH [00:29:00] This side… I’ll go Coronation Street, I don’t know. I don’t know. But we’ll see where that goes.
ANDREW MCLEAN [00:29:09] Dire, dire. British soap operas that was and now my final question. Private jet or luxury yacht?
KEVIN MCGRATH [00:29:17] Oooh. A very good one. I’ll go private jet on that one. I like being on the water. I don’t want to be out there for, like, too long. Like like maybe skiing close to the shoreline and stuff like that. But private jet can get you to a lot of cool places. I don’t need to be out in the middle of the water, I don’t need to be out there.
ANDREW MCLEAN [00:29:35] Great answer, great answer. I completely agree with you. Kevin McGrath, you’ve been a fantastic guest. Thank you so much for joining us.
KEVIN MCGRATH [00:29:43] All right. Thank you have a good one, Andy.
ANDREW MCLEAN [00:29:44] Thank you. That was Kevin McGrath, the Chief Technology Officer at Spot, talking to us about Spot, the things to do, the Red Hat announcement and public Clouds and they’re kind of background story, which was incredibly interesting. The time is now coming up to 1:00 pm on the 1st of June. You have been watching The Andy Show. Make sure you check us out weekdays on disruptive.live and using our #DisruptiveLIVE. But until tomorrow I’ll see you soon.