Want to run your own Amazon 'region'? Stratoscale shows you how

Q&A: The CEO of Stratoscale makes the case for his company’s datacenter OS replacing VMware and OpenStack, thus transforming your servers into an AWS region

Want to run your own Amazon 'region'? Stratoscale shows you how

Stratoscale is a small company with a very big ambition: to turn your datacenter into an Amazon Web Services (AWS) region. Forget OpenStack, forget VMware. Stratoscale aims to help IT shops get beyond device-level virtualization and deliver the same app-friendly building blocks AWS provides. In the process, the company promises to cut the cost of operating datacenters by more than 80 percent.

Founder and CEO Ariel Maislos, who cashed in big in selling an earlier flash memory startup to Apple, says CIOs don’t want to build out bigger VMware-based datacenters. Instead, they want to build Amazon-like datacenters, and Stratoscale has the best solution for those hybrid public/private AWS ambitions. In this installment of the IDG CEO Interview Series, Maislos spoke with Chief Content Officer John Gallant about why longtime VMware customers would gamble on his emerging company and exactly what it means to turn your datacenter into an AWS region using what is essentially Stratoscale’s datacenter operating system. Maislos also talked about why OpenStack – which he dubbed a ‘nightmare’ – isn’t the answer for the dynamic datacenter.

IDGE: Why was Stratoscale founded? What problem did you set out to solve?

Maislos: I founded the company because I was bored, and I needed something to do, and it’s the genesis of a good story. I’m an entrepreneur. I had a few startups. I was very fortunate in my past life and [had] a very large, successful business.

Before Stratoscale I founded a company called Anobit, which I ended up selling to Apple and setting up Apple’s flash storage development team. I headed that for a year before leaving Apple. At the time, Anobit was developing flash storage for enterprise, as well as flash for embedded technologies, the iPhone, and such.

Obviously, at Apple we worked more closely on the device side and not on the enterprise side, but I was always curious about datacenter. The datacenter to me seemed like an amazing piece of technology: very complex, very large, a lot of moving parts, fascinating.

I’m a technologist at heart. I love solving complex problems, and I was attracted to that space looking for something interesting to do. I was always working on distributed platforms and applying them on the storage side building scale-out architectures for very fast flash storage, and I was thinking about applying the same technique for the different areas of the software stack that exist in the datacenter. Looking at virtualization technologies or compute services, looking at the networking services as well as the storage component as well. Looking at all three components, something that’s interesting to solve in a combined manner as opposed to building different pieces that the customer then needs to put together himself. Taking this holistic approach of everything scale-out led us to Stratoscale.

IDGE: When did you found the company?

Maislos: We founded it in December 2013.

IDGE: When did you bring your first product to market?

Maislos: About a year ago. We had version one out when the company was 2.5 years old. We are firm believers in the iterative design process, rapid innovation. We were sitting in front of a customer working on real problems. The faster you understand what the real problem is, the better you are able to resolve it. That led us to a relatively quick iteration where we had version one out a year ago, and now we’re releasing version three.

IDGE: Let’s talk about the products. I notice that you break down the product line into compute, object storage, block storage, things like that. What do these all do?

Maislos: The opportunity that people are seeing is the conversion to cloud. Cloud represents the ability to move very quickly and to build your application with building blocks that make sense.

When looking at the traditional datacenter environment, the building blocks are relatively low-level. You have servers, storage, switches, routers, ports. With virtualization, you get the virtual aspect of the same entities that you can fool around with.

When you’re looking at a cloud, the building blocks are completely different. It’s no longer about low-level stuff but about pieces of the applications that you need in order to solve the true problem you’re trying to solve. They’re not valuable in and of themselves, and you would like to consume them in a manner that saves you time and effort and allows you to focus on the real business problem you’re trying to solve.

That part is exemplified by Amazon, their AWS cloud, and the huge catalog of services they are offering on that platform starting from basic services for compute, data storage, and connectivity, then building more and more on top of that. Our vision is to offer that experience in an enterprise environment, in an enterprise datacenter, and allow them to use the same methodology, the same building blocks when they’re going to build their application on-prem.

IDGE: In essence you’re replicating a set of services that someone would consume from AWS in the enterprise datacenter.

Maislos: Correct. Our evolution as a company has been to get to that stage and build that out. Version one for us was getting some fundamental stuff working. Version two was about stabilizing a platform that is built in a scale-out manner that gives us core services, and in version three that we’re releasing now, we’re starting to add on top of it the different services that live up to that promise.

IDGE: How does somebody build to this? How do they deploy it within their enterprise datacenter? Walk me through that.

Maislos: In an enterprise datacenter you have hardware: servers, switches, storage appliances. That’s a very big capital investment, and we allow you to take advantage of that investment and continue to use it.

Our software runs on top of the existing servers. We’re supporting Intel servers and transform them into capacities that you can consume in a cloud environment. You deploy our software on the bare metal. It’s scale-out architecture, so you deploy the same software on all the machines. They discover each other across the network. They automatically cluster and instantly you have an AWS-compatible region in your datacenter. The entire process can take as little as, say, 45 minutes in a normal scenario – sometimes 48 hours when it’s more complex on a larger scale.

IDGE: I assume the other big advantage of this, in addition to making your own enterprise datacenter more flexible, is that it makes it much easier to work in a hybrid environment with AWS.

Maislos: Exactly. When we’re looking at what our customers are doing and the places they’re going, it seems that the AWS way of integrating data with this ecosystem is clearly winning in their hearts and minds. They’re developing cloud-native applications using these building blocks that are available in the public cloud. If they want to achieve hybrid, they need to be able to deploy the same applications on-prem.

Hybrid is not about taking the legacy environment and extending it into the public domain. We think it’s the opposite. It is being able to take the public environment – which is the growing environment, which is the winning environment – and extending it to your private datacenter and being able to apply the same methodologies in both places. It’s as easy as pointing to a different IP address and bringing the same devops that you’re running in Site A in your private region or in the public region.

IDGE: Does this mean you don’t use other virtualization technologies from VMware or SDN technology from Cisco? What does this do away with?

Maislos: Networking functionality is provided by your routers, switches, and network appliances, and those remain in place. You continue to use them as-is. There is software for orchestrating and managing these network boxes, for lack of a better word, and your software still remains in place. Cisco ACI and things like that would remain in place. The software that you have been using to run your applications and manage the storage, things like VMware or OpenStack or other proprietary stacks, those we would be replacing with our software that would mimic an AWS region in your datacenter.

Forward-looking CIOs are not interested in using VMware vCenter to manage AWS. Instead they are looking to take AWS-native applications and run them in their own enterprise datacenter.

IDGE: People have been using the VMware products, for example, for a long time, and it seems like it’s a big leap from what I’ve become accustomed to and what I have a big investment in to taking a risk in working with a smaller company like yours. Why someone would do that?

Maislos: Excellent question. We don’t think that is the primary motivation. To a large extent, CIOs want to replace their existing VMware way of doing things with Amazon. That’s the decision they’re making. At the corporate level, at the board, at the strategy level, they’re saying: Let’s move to the cloud.

Now they realize, of course, that a lot of applications can’t make that leap, and their existing investment in infrastructure is going to be left behind. We’re giving them that bridge, that ability to make the transition in an easier manner. We’re not coming in directly saying: Hey, we found a replacement for VMware for you. Customers are making that decision by themselves, and we’re helping them do it with their own internal environment in a way that’s compatible with their strategic decision to move to the public cloud.

IDGE: In a previous answer, you made it sound like a pretty straightforward process to do this, but what are the challenges to doing this? Are there some applications for which this is not applicable? What are the obstacles that people would face in making this move?

Maislos: I think that transition to a cloud is never a one-stop, click-of-a-button thing. We’re making their life very easy in deploying a cloud in their environment that’s compatible with the public cloud. That’s a huge leap from what’s available. I’m hearing horror stories about OpenStack deployments gone bad and VMware-based clouds that are taking years to deploy. We transform datacenters into a cloud in hours or days, in the worst case, and we’re doing it as a product, we don’t charge for special services for this. This is not a project. We’re not looking to make your life complicated by giving you tons of options that don’t make sense but allow us to bill you hours. We’re literally making this work out of the box. That’s the first step.

The second step is how can you deal with your applications and data. In transition to the cloud, sometimes people call it lift and shift. It’s about taking the data as-is, taking the application as-is with minimum change and putting them in a new environment that’s friendlier to manage. We give tools that allow you a seamless transition from Hyper-V environments and from VMware into a cloud-native approach that is AWS compatible. We have the tools to allow you to do that that are built in.

IDGE: I want to make sure, because you use the term AWS region, and I want to make sure readers understand exactly what an AWS region is.

Maislos: In the AWS world you have regions like US-East or US-West or London, and the idea is to create a mini region in your own datacenter where you have your EC2 instances running and so on.

IDGE: What is the set of services that you currently offer? What are the things that you’ve replicated from the AWS environment?

Maislos: We focus on the core services, starting with offering compute services, EC2 for virtualization and the equivalent of ECS, the container service where we’ve implemented a Kubernetes-as-a-service engine that allows us to provide those services. That’s on the compute side. On the storage side we offer block storage services, EBS in the AWS jargon, and object storage known as S3 and lastly, networking services in the VPC family of features. We also have a built-in, multisize VR capability, and we have an application catalog or an image marketplace, to be more exact, built into the infrastructure as well.

IDGE: What’s been your success so far? How many customers do you have?

Maislos: We have more than 20 global customers right now. We’ve been very fortunate working with very early adopters that have been with us right from the beginning, helping us steer the direction of the product. We’ve been iterating with them over the past year. We’ve recently beefed up our sales organization and starting hiring customer-facing teams on a global basis and now we’re beginning to grow.

IDGE: Can you give me a customer case here? Who is using this, and what are they doing with it? More important, what benefits have they achieved?

Maislos: We’ve had, for example, what you may call an IT-as-a-service team that is providing services to local municipalities and government offices out of Europe. They are helping them deploy applications and services. That’s a clear example where they can do a lot of development in the public cloud but because of privacy concerns and the nature of the data, it has to be in a protected, government-owned facility, and you’re getting into a skill-based deployment paradigm. That’s one case that we saw.

1 2 Page 1
Page 1 of 2
InfoWorld Technology of the Year Awards 2023. Now open for entries!