Demystifying Cloud Resources: VM/GPU as-a-Service 101 - Rob Foster

Introduction

So I'm going to go through a few things and the most important one is that free money bit at the end and I promise we'll get there.

About The Speaker

So hello, I'm Rob Foster and I'm the infrastructure manager at NextGen Cloud. I've previously worked at places like Rackspace, Accenture and Johnson & Johnson. I've been in IT for longer than I probably should admit, but I've been around.

If people want to get in touch, then LinkedIn up there, feel free. And I've been in London today at our London office. My office manager is missing me, who is Charlie the cat, who makes more sure that I stay honest.

And my slides have disappeared. And yeah, he's on Instagram. He's much more photogenic than I am. If you see a cat, it's because I prefer his picture to me.

Introduction to Next Gen Cloud

So who are Next Gen Cloud? So we're a European IaaS, so Infrastructure as a Service. We're specializing in HPC and GPUs.

1The reason that we're doing this is because we feel that there's a whole lot of market out there that can't get access. And if you can't get access, you can't get started. And then that stops a lot of the innovation.

So we really are quite proud that we make our platform available to everyone if they want a GPU for an hour, they want it for a month, they want it for a year. it doesn't, there shouldn't be a barrier to getting that innovation going.

And we see a lot of entrepreneurs, we see a lot of startups, we see a lot of ideas, and we're really proud of being part of that. And our founders have got a mission, and that's to, they call it democratizing accessibility. And that basically means just that.

It's a making sure that if someone wants to access, they don't have to go into a five-year deal with AWS, they don't have to go and spend millions of dollars. They can get access to that resource straight away

Next Gen Cloud and our main product is HyperStack. You'll see there's a couple of different marketing names, but the company's Next Gen Cloud, the product is HyperStack.

Product Demonstration

So I'm going to go through a demo in a minute. I'm going to go through the GUI and just show how easy it is to get started. We'll go through the API and that's what I'm really proud of because anything you can do through the GUI, you can do through the API. If you can do it through the API, you can script it and you can really make it repeatable.

So you can get in, fire something up, test it, kill it. So it's only cost a couple of cents rather than lots of time to actually get that resource going. But as I say, everything's in the GUI for people who want to click. Everything's in the API for people who are more comfortable with doing that.

So, let me flick over to a bit of a demo. Now, yeah, it's in the desktop, which is great. So, apologies if you can't hear me, I've been told I've got a very soft voice, but, so yeah.

So we're trying to keep things simple and we're trying to keep things clean and there's not a whole lot of substance. If you're used to things like the AWS console, you'll be used to thousands of buttons. Well, that's not really us. What we prefer to do is to... Keep it simple. We want it simple, easy. We don't want to bloat it with buttons.

Every time we're making a GUI change or an interface change, we're trying to put the customer at the center of it because not everyone's an IT expert. Some people have got that great idea. They've got that thing they want to do. They don't want to be hindered by having to be a coding expert or someone who needs to know their way around portals.

So once you're registered and at the end, we're giving away some free credits for people to come and take a look. It literally is as simple as coming in and going, I want to deploy a new virtual machine. I want to call it CAT1. I've created a test environment. I'm going to create whichever flavor I want. And again, there's lots of different cards around depending on what workloads you've got. I'm not going to go into the differences for those, but in short, different workloads, there's different compute resources that are better and might do it more efficiently.

We can select that and we can hit deploy. Depending on load, depending on availability, that can take 30 seconds. It might take a minute. Some of the bigger machines, they can take longer just because of how much memory they've got. And that's a limitation on that card. It's got to go through and do a memory check before it hands it over. And that might take five, 10 minutes to do, but you're not charged for that. And like I say, that's the bigger machines.

But this right now is in that build queue. It's already going through. And as I say, I'm fairly confident that within a minute, that's going to be built and that's ready to use, can come in, do what we need, kill it at the end. And again, I'll come onto that in a bit as well. So that's already active. So that was what, under a minute to get a machine that we could log into, load up our data sets, load up our application, test it out, costing literally pennies. And that's using the latest, greatest GPUs.

API and Automation Features

Now, if you want to go through the API, the API is cool and I love the API because for me that's where the magic really happens. I've already prepared just a short piece of code which our documentation is all there to help people through it. And if I just paste that in, that's now in the build queue. I could have done that as a script at five to nine in the morning, create the machine, so by the time I come in with my cup of coffee, it's there, run my workload, delete it. I could run that automatically to where overnight it builds it, runs tool, deletes it, and I've just got the output in the morning. It really makes it nice and easy just to be able to do stuff without making it time consuming and difficult.

And that's, like I say, what we are really proud of is that we want to make it easy.

I'm just going to get the VM ID for that was 34083. 34083. And the reason I paste it is I don't trust myself to type.

And straight away, unauthorized because of course, why isn't it? My token's timed out. Again, curse of the demo.

Demos never go right ever since the earliest days. So just needed to re-authenticate, I think.

Three, four, eight, three. still telling me unauthorized typical i'm going to come into where we can see this um it's already active so it's uh it's got my test server name so that was again under a minute to come through and do that and nice and quick and easy now it doesn't always work it might give an error might not be stock available because it is a cloud that people can consume and we try and feed that back you know it's that's that's what makes it so great is that we try and give you control

So let me come out of that because I was gonna go more into this. I don't think that's what I'm gonna do today. And this is where Carmen's probably panicking a bit about what I'm gonna talk about next.

So let me come back in to here.

Charitable Initiatives and Sepsis Awareness

Shameless plug alerts. So one thing that I'm really proud of is that I do a lot of stuff outside of work.

And one thing I'm doing later in, well, early in May, I'm gonna be doing 106 kilometer walk for sepsis. Now, if you don't know what sepsis is, i urge you to look it up and and learn a bit about it it's it's a it's a horrible illness a good friend of mine ended up almost dying and that was really how i found out about it and it is something that is close to my heart i try and always have a a charitable thing that i look at because at the end of the day we're all human and as much as we've got it stuff that we're doing it's really important to think beyond that and think of the community so if i can ask you to take one thing away from this go and take two minutes to learn about sepsis Because if you learn the symptoms, you could save the life of somebody that you care about.

And now we'll return to the regular schedule program.

Security Considerations for Entrepreneurs

So the next thing I wanted to say was some back to basics, and this is moving away from the technical bit, but something that I've realized over the past few months, talking to customers, listening to people who have got great ideas, is a lot of times they're so focused on the idea that they actually are forgetting some of the things that are really important.

So I want to talk a bit about security because over the past few months we have seen a massive surge in attacks against our customers. Now these a lot of times are the entrepreneurs are coming up and their application's not working and we look into it and actually they're being attacked. We've got one particular customer within five minutes of him launching his environment, he was being attacked.

and it was causing him to waste time, waste money, thinking that he was causing a problem. No, there was someone out there, for whatever reason, the chosen, they were going to attack him. So things like DDoS, if you've not heard the term, distributed denial of service, it's basically when other people on the internet are attacking you in one of a number of different ways.

Now, it can be random or it can be targeted. Now, we've seen it where people are just out there and they might just be targeting an IP address, but in this particular customer's case, someone had decided they were going to target him. Now, that might have been a competitor, it might have been someone with a competing idea,

every time he was trying to launch up and test his application, it was being attacked. Now we helped him mitigate that, but it's something that having the knowledge and knowing that that is a real case. So it's not just about having a great idea. Unfortunately, you've got to be aware of things like this.

Things like phishing and social engineering. Every day we have people contacting us going, hello, my name is Fred and I want to do something. Well, it's not Fred. We know Fred and you're not Fred. You've got to be really careful about it.

And again, if you've got something, if you've got data, if you've got an idea, if you've got something intellectual property, there is going to be someone out there who is going to be trying to take that away and steal it from you. So something really to be aware of. And yeah, humans are the weakest link in the chain.

I think it was recently T-Mobile had a massive outage that they put down to someone rang up for a password reset. They reset the password that gave an attacker access to the systems and it took down their network for days. So this isn't just something affecting small people. This is affecting the biggest companies out there.

And credential spamming. So people are out there actually going through literally trying a whole dictionary attack of passwords. You can't just rely on your mother's maiden name and a birth year. People are going to find it out. People are going to spam it. They're going to break in, break what you're doing, and it's going to slow you down.

There are things we can do though. So, you know, plan to be secure by default. Don't think, yeah, I'm going to get up the great idea and then I'm going to bolt security on later on. It never works out. Think about it from the start.

Make sure that it's part of what you're doing so that when you come up with that great product, that great idea, it's already secure. Passwords are bad. End of. Certificates, pass phrases, there's other things to do. But if you are using a password, someone can guess that password.

lockdown access to what's needed. Again, we see a lot of people coming in and going next, next, next, and they end up with a web server that's open that they didn't really need, but it's there because it was part of the base install. Think about what you're doing. And again, it comes back to that being secure by default.

Log what you're doing. If you're expecting 1,000 connections per second and you suddenly get 10,000, why is that? That might be a good thing. That might be a bad thing. That could be an indicator that someone's trying to do something. So don't rely on things always working.

And again, respond to it. It's one of those think about what's going on and don't just rely on the fact that crossed fingers is not a strategy. And know your market. This comes back to the customer that I was talking about. It turns out that what he's doing, there's a lot of competition and he thinks it was a competitor that was trying to slow him down so that in the market he was in, as fast as his application could respond, if someone else's application responded quicker, they'd end up getting the business. It was all to do with blockchain and being able to authenticate transactions. So the difference of a couple of microseconds made a difference between him getting business and someone else getting business. Once he understood that, we could actually do something about it and help solve that problem.

Operational Health and Cost Management

So then something else to think about is operational health. So cloud services do get expensive fast. Amazon will tell you you're gonna save a million dollars. Come to us, you're gonna save a million dollars.

Well, you're only gonna save a million dollars if you're spending more than a million dollars. So keep an eye on it. You've got to have a plan. You've got to have a budget.

You've got to think about it in advance because otherwise you're going to run out of money. It's going to cost you more at that most critical time. And it sounds simple, but I've seen so many people overspend and then that completely destroys funding, business models, ideas.

1Pets versus cattle, now this is getting a bit, I've got a vegan friend who gets very offended when I talk about pets versus cattle. So we came up with oak trees versus cornfields, just to make everybody inclusive. In short, one you take care of, the oak tree you're gonna plant, you're gonna look after, you're gonna have it for 100 years.

The cornfield you're gonna raise, you're gonna grow, you're gonna cut it down and regrow it again. Don't build things and keep them if you don't need them. Get rid of them. They're going to cost money all the time they're running.

So make it so you can kill it and replace it when you need it. Comes back to what I was saying about the API. Run that machine when you need it. Get rid of it. Limit your costs. Limit your exposure. And then you can be in a much healthier position.

Human Touch in Technology

And then finally, don't forget the human touch. We're all in IT here, we're all looking at AI, but at the end of the day, we're the humans. Hopefully, we're still in charge. Have the empathy because the systems don't.

When you're designing, when you're coming up with ideas, it's not all about the bits and the bytes. You're doing something that is affecting the lives of people.

Last week we saw a talk from an entrepreneur who's doing a vision tool that's helping people that are visually impaired. It's fantastic. I mean, that's a really empowering thing to think of. We've all got that power within us.

In a world where you can be anything, be kind. There's so much hatred out there. Let's come together. Let's use AI and technology to actually make a change rather than just letting it be part of the problem.

And let's stay in control. I don't want to be bowing down to a robot overlord anytime soon. So let's do it responsibly.

And celebrate the wins. There are going to be times where everything's going right. Fantastic. There are going to be times where it feels like everything's going wrong. It isn't. Don't let that get you down. There will be failures. There will be times where you think, yep, the world's against me. No, no, it's not. It's just things that go on. Don't beat yourself up. You know, get up, brush yourself off, get on with it, because it's one of those things that we're resilient.

Special Offer: $100 Free Credit

Now, the free money. So we're offering $100 free credit to use the platform. There's no strings. You don't need to register with a credit card.

There's an email. If you've got a Google address, I think you can get through in four clicks. It literally is $100 credit, fire up resources, use it, try it, use it to spin up an idea.

We want to get our name out there. We want to get people using platforms like this to do more.

Feedback and Final Thoughts

And this leads me into, we want some final thoughts. So let us know what's good, what's bad, what he's not doing. If you come in and go, feature X, you haven't got feature X, you need feature X, let us know.

We're trying to keep things simple, but equally, we want this to be a platform that works for the people. There's no point in working for one person if there's actually something else we could do differently. You know, don't be afraid to use your voice.

You know, I'd rather hear 100 people saying this is rubbish and it needs to be different than not hear it and just, well, hear the sound of their footsteps leaving. So please, please, please have a look. Feedback. Let us know what you think. Let us know what's good. Let us know what's bad.

and carry on being awesome. We're living in a brilliant time. If I look at the timeline, what a time to be alive. I think now we really are living in an industrial revolution of a technical stage. I think that if we look back to our parents and our grandparents' generations, I'm sure they had stuff going on, but nothing like this. This is such a brilliant time to be here. So yeah, so thank you.

Thoughts, questions, ideas. I really do want to hear from people who've got stuff to say now, later on. But yeah, thank you very much.

Finished reading?