Using agile development to build storage for SMEs

European storage vendor Syneto has its software development in Romania, its commercial headquarters in Italy, and its manufacturing in Belgium. The company puts agile development at the heart of its product development.

ZDNet caught up with its CEO Vadim Comanescu, to find out if he and his partners aim to go on doing it that way.

ZDNet: I understand that your two most senior managers -- yourself, and the other cofounder, COO Marco Lorenzi -- have swapped roles: he was CEO and you were COO.

Comanescu: We started out in 2000, and between 2000 and 2009 we specialised in network appliances. We had our own operating system for networking and network security.

In 2009, we decided to shift the focus of the company to storage. We are a European company and a joint venture between Romania and Italy, our commercial headquarters is also in Italy and our R&D is in Romania.

Right now, from a software development perspective, Romania is a very popular place. Our manufacturing facilities are currently in Belgium. You can see that we are a truly European company divided between three countries.

On the role-swapping, we wanted to move the company on and continue with our expansion. So, because of the language barriers and other things, we decided to swap roles.

But really, in the last few years I have been acting as CEO anyway. I have been dedicated to the expansion in Europe and further abroad. We have a small presence in the US at the moment and in various parts of the Middle East too.

I was the initial starter of the storage project and the development of our storage operation, so you can see it made sense.

Actually it is quite a funny story, because I started in the company doing technical support before I became lead engineer. As lead engineer I created the bits and pieces of our storage operating system, Syneto Storage OS. And from there I shifted my focus into operations and development. Now I have taken over the role of CEO.

You build a technical company like this it is not enough to be a good marketer, you need a strong technical background as well.

It's a good European story too because there are not too many European companies who are developing a storage operating system, and not many who are building what we call software-defined appliances.

How would you describe your company's point of difference?

Whether you are a mature company or a startup, you need to understand who your customers are. When we started, we did an analogy: as a startup, you can choose whether you want to be a deer hunter or an elephant hunter. Everybody wants to be an elephant hunter, but it can be better to go after deer. Our target is SMEs in Europe -- 99 percent of Europe is made up of SMEs, and we believe there is a way that we can make a difference there.

When we started building a system, we knew that most of Europe is used to traditional storage systems. We decided to make our system a more enterprise and feature based solution. The world is moving towards software and that is where the intelligence is going too.

So, one way to do this is through software designed appliances based on general purpose hardware with all of the intelligence moved into the software stack.

The second thing we discovered, about three years ago, is that hyper-converged movement and software actually fits better in the SMB space than in large installations.

So, from April 2015 we started deploying a lot of hyper-converged, all-in-one, ready-to-go appliances for SMBs. That is being very successful for us.

Then if you want to build a business based on appealing to small companies you need to specialise, so we have been pursuing vertical markets. One is CAD. There are about 30,000 studios who need storage solutions to manage their projects.

Then there are areas like hybrid storage arrays and flash storage, then on the NAS side we have seen a lot of CAD companies that use Apple computers with NAS.

What we find with the PC architecture is that users cannot buy a reliable, enterprise-ready infrastructure that sits well in their Mac environment. Apple doesn't sell it, so we have built a system that can integrate and do things like index all of the files from the storage side. With it, you can search through six or seven million files in something like 100 million milliseconds.

We have an implementation up and running with one of the biggest publishers in Italy. That is an area where we can bring value and something different.

There are just two ways to build a hyper-converged solution: you run your OS inside the hypervisor or the hypervisor runs inside your OS. Most of our customers run VMware as a VSA [virtual storage appliance] and most of the appliances we ship are targeted at hyper-convergence.

They are running as a VSA targeted at the convergence part and bringing all of the functionality like disaster recovery, replication etc. We aim to bring a lot of pluses on the storage side such as in-line compression, incremental replication, etc. All of those things that a lot of our customers want.

Hyper-convergence is a popular buzz-phrase. What's your take on that?

It's like software-defined storage. What does it mean? All storage has software running on top of it. My take is that it is the simple unification of compute network and some storage in a single box.

For an appliance, you understand that when you connect and operate such a device, without any special intervention, you can easily or automatically deploy an application.

From my perspective, the application is the fundamental part here. Without applications you don't need the infrastructure, so when a client connects to such a device he can write, say, new VMs for the application without having to provision whatever storage is underneath and without constantly having to connect complete new infrastructures, connecting to switches, or whatever.

Another area of complexity are silos where you have one entry point. You have this box and if something doesn't work, who do you call? And how good are they at fixing it?

So what we want to do is Agile, Agile infrastructure deployment. But we don't have our own infrastructure to put over each deployment. We asked, what is the most common interface which people are using now and everybody knew and was comfortable with.

Q: You're a small company and I am assuming you are venture capital funded?

At the moment, no. This is what I think is the beauty of this story -- we have actually arrived at this point by ourselves. Most companies would start with a big slice of venture capital funding and then spend five years putting a product together and coming to the market and saying, 'here it is'.

What we did was spend about 10 months developing and putting together a storage system. Then we, the founders, jumped into a car and went around trying to sell the product.

You know, everybody talks about big venture capital funding to get going, and I have nothing against that, but there are actually a lot of companies out there who do good things and bring value to people and value to the customer.

They don't get the same attention as the guys spending €300m but we have grown organically and by 50 percent year-over-year. I know that, at some point, we might need VC funding to accelerate the growth, but for now we are doing it all ourselves.

And you know, that means that when we go out, we can see what makes the difference and what is useful for people and what is not.

I tell this story because I am quite proud of it.

Q: So, from what I see of your growth, you started with two different pieces of software and as you have grown you have produced another component, then another, and so on?

Yes, yes, and yes. I was a programmer myself so I love software, I love technology, and I love all of this and when we started the Syneto story it was a project.

That was in 1999 and we started using Agile methods. After a year, we had 10,000 automated sets that ran on each commit that was going onto the code base of Syneto storage. That was only the beginning.

We found out that Agile was the only sustainable way to build it without creating a mess every time we tried to update it or introduced new features.

So, in our own small way, we are trying to change the way in which software is developed. We want to do continuous sustainable development and deployment, and we want to come out with six features in one week. We actually operate in one-week iterations.

Of course, people in this space are not ready for that just now. It is probably the understandable paranoia of the typical manager, but I do believe that that is the future.

The other thing that makes this work is our own structure. We work in small teams of software engineers who work in an Agile way.

A lot of people don't see that. They don't see the advantage but we had the privilege of actually doing things in an Agile way.

Further reading

• Australian government tenders for five qualified 'Agile coaches'
• Do we need 'industrial-strength' Agile computing?
• Let go of certainty, start using your brain: How to make the most of agile development
• Elymentz messaging app helps simplify SME's telephony solutions
• SAP reshuffles product, executive decks to better target smaller companies
• How to apply Agile practices with your non-tech team or business (TechRepublic)