Exclusive Interview With Boyan Ivanov, CEO, StorPool

Boyan Ivanov, 38, is CEO and co-founder of StorPool, a company he launched in 2011. Before that, he worked at Stedex, webEndians, ABN Amro bank, DeltaCom 9 and Alpha Finance Holdings. Outside of working, he loves traveling and biking.

StorageNewsletter.com: Give us a snapshot of StorPool as of today, especially around your financial, employees, offices?
Ivanov: StorPool was started in the very end of 2011 with the goal to build a storage software which is able to replace a high-end SAN or AFA. Mind that this is before buzzwords such as SDS) and Software-Defined Datacenters (SDDC) even existed.

We as founders were serial entrepreneurs and the core tech team was extremely technical. Our tech guys have built service providers from scratch in a number of countries in Europe and Asia and knew a lot about the entire IT infrastructure stack – from rack design, cooling, power, servers, storage and networks, up to the application layer. It is this multi-disciplinary knowledge that allowed us to create a unique product where others have designed solutions only by using their storage-related background and innovated on 1 axis, when you can actually innovate on several axis, all delivering a radically better storage product.

Today we are an extremely efficient team of about 40 people with multi-million dollar revenue. We are still headquartered in Sofia, Bulgaria, while our customers are on 4 continents in some 30 countries.

[According to crunchbase, the company got $200,000 in seed funding in 2012 and unknown series A investment in 2014]

How did Covid-19 impact your operations?
Covid has accelerated adoption of IT in general. And StorPool, as a piece of the backbone of IT infrastructure has increased its business with 80% in 2020, due to strong demand for a best-in-class storage product.

We see strong demand continuing in 2021, even though the small picture is dynamic – for example we’ve seen some projects get delayed, frozen, or cancelled. While other companies launch new digital transformation initiatives to stay relevant or to quickly ramp-up their IT capabilities and adapt to the new normal of working from home and having distributed teams. Or to satisfy the significant increase in the demand of their online products and services, which is the case of e-commerce for example.

Also – now physical borders disappeared, which also acts as an accelerator for our growth. Before many companies preferred to have sales reps visit them in their offices. So geography mattered more. Now everything is done online. Whether you’re based out of San Francisco, New York, Dublin, Sofia or Delhi – doesn’t really matter anymore. It’s now all in the same place in a way, because it’s in Zoom or other video conference tool, in an aligned global time-zone.

What is the genesis of the company? What were the triggers to start StorPool? Give us some context.
In 2011 what became the origin of StorPool was another product company, building what is now referred to software-defined networking. Back then SDN was not as established, and as part of our customer validation process we got overwhelming feedback that the most problematic piece for large data storage users, such as banks and service providers was storage. SANs were too slow, complex, expensive and inflexible.

So we decided to take a “blank sheet of paper” approach, discarting everything we knew about how storage was done -and to design the storage solution of the 21-st century. It was apparent to the team that it has to be software, running on standard servers. There was nothing at that point that required single-purpose, specialized storage devices. This is now obvious, but was counter-intuitive back then. I recall our first customer chats when showcasing the early StorPool versions. People would not believe that a piece of software can be faster and more reliable than a high-end SAN.

While there were some first-generation storage software products back then, they were not really able to replace a proser SAN. They were the cheap and sub-par alternative. Again this is before SDS was a thing and before solutions like VMware vSAN even existed. 

And there we were, showing a solution which was faster than what people believe was possible, at a fraction of the cost.

I recall our demo with a VP at one of the biggest virtualization companies back then – “what you are showing me here cannot be real. It is faster than what the physical drives can deliver. And we tried to build something like this, but failed. How can you do that?” Well, people did not seem to grasp the power of a proper distributed storage system, build for speed and scale. Now people do not want to buy a storage array anymore, they look for SDS.

Founded almost 10 years ago, you have raised only a ridiculous amount, how did you bootstrap the company? Is it founders and/or friends money? Or even from other sources like government?
We’re a company with a very different DNA. We believe in being the best at what we do and being extremely efficient. We believe in doing the right things, where others look for shortcuts. This translates to how we built StorPool as a company – very efficiently, with minimal funding.

We started with our own money and got only a couple of VC seed rounds. Contrast this to our US competitors which raise huge amounts of money to create beautiful slideware, but poor products or our other European peers, who get government funding, which we deem a wrong tool.

We set out to build the best product in this domain in the world, all on business principles that are sound. We’ve built the best block-storage software for demanding customers – in terms of speed, reliability, efficiency, scalability, management.

This leads to StorPool having real paying customers, growth and revenues, which can fund the company. We’ve been growing organically. And even though not as well-known, for the last several years we are one of the handful profitable storage players on the market.

Does the mission evolve with OpenStack wave and SDS market dynamic?
These are small details in the grand scheme of things. The large picture is that the entire global economy is changing. Business demands are changing. All due to the fact that any business now is becoming essentially an IT company, with a product or service wrapping.

Think of a bank for example – in the old day their main assets were branch offices and front office clerks. Now their customers interact with their online banking – be it through the website or through a mobile app. It’s all digital and the core asset now becomes the digital infrastructure, not the physical layer.

This shift requires much faster IT infrastructure & storage. It has to be reliable, seamlessly scalable, able to service millions of concurrent requests, at extremely low latency, i.e. speeds. And this applies pretty much for any “traditional”, non-IT business – they all become increasingly digitized. This is what “digital transformation” is.

This master trend causes many tectonic shifts and the technologies/approaches you referred to are just some of the waves, caused by the tectonic shift.

For example:

• Because of the sheer scale, modern IT systems have to be API-driven, self-service, multi-tenant, very fast. “Cloud-native” if you’d like.
• The economic buyer is changing – because now everything is becoming IT – but 95% of businesses are not IT companies – they need someone to help them to transform. Thus they go to public cloud, but also they go to their local MSPs  and SIs and hosting and/or SaaS companies, which become their “external IT”. 10 years ago, the major acquirers of IT were the Fortune 5000 companies. Now the major acquirers of IT are cloud companies in all their forms and shapes – SaaS, PaaS, IaaS, MPS, telco, etc.
• This drives tremendous demand for skilled IT talent. But skilled IT people do not want to go and work for a brick-and-mortar business at the back of the wave. They want to be at the front, working with the most innovative technologies, at scale, doing cool stuff. So they land in innovative companies like StorPool and the others. And because these 95% of businesses cannot attract, utilize and keep the best talent – they have no choice, but to shift to specialized IT providers – like public or managed private cloud builders, SaaS, etc.
• Everything is shifting to as-a-service model and the entire economy is consuming IT and pretty much everything else on a recurring cost/revenue basis.

On the technical side, the above drivers translate, as follows:

• The digital transformation cannot be done with the technologies of the past – hypervisors, SANs, networking from two- and three-letter tier-1 vendors. Their technologies were developed for a different user, at a different scale, with different needs. So the business of today requires new tools to meet their new requirements.
• Thus we see the trend of severe commoditization of IT – pervasive usage of Linux, KVM, open-source technologies, i.e. cloud native. They are scalable, automated, cost efficient, etc. For example: 50%+ of workloads running on Microsoft Azure are Linux based and this is going to increase.
• Just try and ask “How much VMware or NetApp does Amazon AWS use?” – it’s a ridiculous question, isn’t it? …That’s what’s happening to the players of yesterday.

What is your market or to be precise do you target some specific use cases and users profiles and needs?
StorPool has been born, designed and optimized as a solution for the cloud era. We call it “New- Age” or “Modern IT”. It’s all the things I mentioned above, all the things that businesses need today and are not delivered by the incumbents of yesterday. StorPool is extremely fast, seamlessly scalable, agile & flexible. It’s best-in-class software, fluid and malleable, yet extremely robust. It does not impose lock-in. It automates and optimizes operations.

All this means that we naturally appeal to the companies that build New-Age, cloud-native IT systems, but outside of the 4 or 5 big public clouds. So our typical users are MSP, telco, cloud and hosting, SaaS, e-commerce, SIs, sometimes hardware companies, which are looking for a solution to OEM.

Some of the typical use case we serve are:

• Providing the storage layer for demanding virtualized, containerized or bare metal infrastructure, at scale – think of databases, OLTP, VDI, demanding web/SaaS applications.
• Software stack consolidation – we are one of the most widely integrated solutions in the New-Age stacks – we can have a single storage platform, which can service multiple workloads – say a rack of VMware servers, some Hyper-V, some bare metal, some Kubernetes in a different rack and say OpenStack – all from a single storage system.
• Hardware stack consolidation – larger companies have typically grown over a number of years, oftentimes through acquisitions. Thus they have a horrific variety and complexity of hardware platforms – multiple generations of hardware, from multiple vendors, for multiple use cases – it’s just a nightmare. So they use StorPool to get rid of all this complexity, replace it with a single solution and restore their peace of mind and sanity.
• Hardware lifecycle management – another big reason why companies adopt StorPool is to move to the so-called SDDC or Software-Defined Datacenter design. See, if you have a large datacenter – the hardware maintenance, upgrade and refresh cycles are a huge problem. They are very disruptive, risky, expensive and time consuming. By using StorPool’s powerful data management platform they decouple their data from their hardware. So you have all your precious data “living” in the software layer and you have a “river” of commodity hardware flowing underneath. So all the pain associated with hardware management at scale becomes a thing of the past. We make this problem disappear.

StorPool is a block SDS? Give us some details on the product and especially your differentiators against classic offering?
When we started StorPool we focused on the most demanding use-cases and customers. This meant block storage for critical, tier-1 workloads – databases, large applications and virtualized environments.

Furthermore we wanted to build the best new solution in this category from the ground up. This approach was discarded by most vendors on the market, because it’s so hard, time consuming and expensive. But it’s the only right approach.

Compared to standard SAN or All-Flash offerings we have many advantages:

• No specialized hardware and vendor lock-in
• Using standard servers, which allow you to run in an HCI fashion or stand-alone, in an composable infrastructure fashion.
• Local SSD performance, coming from a highly-available shared storage system
• An Opex, pay-per-use pricing model, no more large Capex upfront payment or low hardware utilization for months and years
• Single storage platform for a wide variety of use cases and workloads
• No painful migrations, hardware generation changes
• Online scalability of a multi-controller, scale-out system
• Advanced data management functionality on par with or better than high-end storage boxes. Etc.

And also against other SDS products?
Compared to the various other SDS product we have numerous advantages, for example:

• some solutions are not scale-out. We scale seamlessly, online.
• others are very slow and cannot deliver NVMe latency for example. StorPool does <0.100ms.
• some SDS solutions are just a “feature play”, not a storage product in themselves. They are an add-on feature to a hypervisor. You cannot use them without the entire stack of that vendor, which imposes vendor lock-in. Whereas StorPool is a true storage solution which can be used with any software stack, although we excel in the modern Linux/KVM and container stacks.
• Some solutions trade speed for features and vice versa. StorPool delivers astonishing speed, while providing a long list of data management features…

And in terms of data services, give us the list of features your product delivers?
We focus on the features that make sense for the New-Age cloud. The most useful and used features are thin provisioning, QoS, multi-tenancy, snapshots and clones, erasure coding (coming), encryption, built in backups & data protection.

Contrast these to the typical features that are pushed by legacy vendors, telling everybody that they need dedup and compression, or else … the world will end. Well, for some use cases these features make a lot of sense, especially for the traditional enterprise. But in a cloud environment these are not the must-have feats, since the data and data usage patterns are very different. Then you need the features which I mentioned above – these are the must have ones.

Also we see people looking for “unified storage” – block, file, object all in one product. This concept is appealing, but elusive. In practice the requirements for the three are conflicting and they need a very different low-level architecture to work well. Thus we see customers usually doing the hard but right thing, which is selecting the best-of-breed products for each of the 3 levels. And at StorPool we excel in block and containers.

What is the split between iSCSI and StorPool client deployment?
The majority use the StorPool client, especially with the deep integrations we have with multiple cloud and container management platforms – OpenStack, CloudStack, OpenNebula, Kubernetes, OnApp, to name a few.

A large and increasing number of deployments are also using iSCSI, in many cases in parallel to the StorPool client, to provide storage to VMware, Windows, Oracle, bare metal and other environments.

You already support NVMe, do you have any plan to support NVMe-oF, and as you already offer iSCSI, I anticipate it would be with TCP?
We already deliver NVM-latency shared storage to the tune of 100 microseconds, through our block driver by doing kernel bypass and other techniques. Our iSCSI speeds are not materially different. So in a way we deliver what NVMe-oF is used for.

Now we are looking into implementing directly NVMe-TCP for wider compatibility.

How do you see Kubernetes growing landscape? Especially with recent acquisitions with Portworx acquired by Pure Storage, Kasten by Veeam or some partnerships such the one between MayaData and DataCore?
First, in my view, Kubernetes, or at least containerization is here to stay. It is becoming a consecutive technical paradigm, much to the scale of the shift from bare metal to virtualized servers. And the drivers for this are the ones I mentioned earlier – business demands are changing, resulting in new types of use-cases and workloads, which require the portability, automation, scalability that containers provide better than previous technologies.

Second, the landscape is still noisy, but will clear over the coming year or so. There was a new crop of vendors that specialized in K8S, some of which you mention. There were some more established storage software companies that entered the domain, like StorPool. And of course there were the legacy SAN/AFA vendors trying to get a piece of the action.

Depending on the use case and and type of the user – companies may go with the first and/or second group. Users would not go with the storage box approach unless they already have the box. If you think about the “cloud native” movement, it’s all about “hardware as code” on standard servers. So SANs/AFAs do not belong in this new world.

On the alternative between latest K8S focused startups and more established SDS, users can choose between:

• the new products, offering more advanced features for Kubernetes, which live natively in the public cloud, but have lower product maturity and sometime no ability to run on-prem or proven SDS solutions, which are production hardened over the years and can be deployed both on and off-premises, but may be less tailored or have some other gaps as to their Kubernetes capabilities.

So it’s really up to the users to decide what they prefer and which they run where.

Your firm jumped into that demand, how do you support Kubernetes?
StorPool was unique in that it is both latest-generation SDS, which can cover multiple IT stacks, yet it’s proven at scale with mission critical customers. In other words we had an extremely strong SDS solution, and extending it to support Kubernetes was yet another IT stack we added support for. Initially we integrated several years back and we’ve re-designed the integration to CSI, once that was introduced in k8s.

People can learn more about our Kubernetes support here, if interested: https://storpool.com/kubernetes

What about cloud, how do your SDS work in such model as you’re not listed on AWS, GCP and Azure marketplace?
Currently we do not run StorPool in public clouds. We are the elastic block storage for on-prem environments, if you will. We’re working on hybrid cloud capabilities though, where companies can send data to public clouds for backup and DR purposes.

How do sell your product? Is it channel only? What about OEM?
We predominantly go to market directly. However we do have a few selected channel partners, who build entire solutions bundling hardware and software for the customers, who prefer that.

We also have some OEM partners.

What is the pricing model? Give us some examples for 100TB, 500TB and 1PB.
We sell solutions, covering different use cases and performance tiers. Thus there is no single answer to your question, it depends on the particular project needs. So we have different pricing for different use cases and payment models.

We typically price per terabyte of data in the system, on a pay-as-you grow basis. This includes the software and our fully managed services &amp; proactive monitoring and issue resolution, which are highly valued by our typical customers, who manage datasets of hundreds or thousands of terabytes. However, we can accommodate more traditional Capex-based models where necessary.

How many customers do you have? What is the split between geos? And between different uses cases? And how many instance of StorPool are running? What is the total capacity in production?
We do not disclose customer numbers, but I can share some other metrics. For example we have customers on 4 continents with about 50% in Europe, 30% in North America and 20% rest of the world. We have tens of petabytes of capacity under management, on thousands of servers.

What can we expect from StorPool in 2021?
Interesting new product capabilities, for example we just release a new version of the software and moved to CI/CD development of the software. Looh at https://storpool.com/news/storpool-storage-v19-2.

Other than that – we’ve onboarded some interesting large new customers, so watch for announcements around customer wins and interesting new use cases.