New DataCore CEO looks to MaxParallel to drive expansion

New DataCore CEO Dave Zabrowski said he is looking to bring the vendor beyond storage by delivering its new MaxParallel workload optimization software across entire data centers.

Zabrowski's career as an executive began at Hewlett Packard, where he worked from 1997 to 2002. He rose to vice president and general manager of the $2.5 billion business, responsible for enterprise servers, storage and PCs. He served as CEO at networking virtualization pioneer Neterion from 2002 to 2009, and then founded and ran Cloud Cruiser, a cloud analytics startup that Hewlett Packard Enterprise acquired in 2017.

Zabrowski joined storage software vendor DataCore this month, replacing founder George Teixeira as CEO. Teixeira has moved into the executive chairman role, working with the management team on strategic initiatives for the company that he founded in 1998. DataCore claims to have more than 10,000 customers using its SANsymphony, Hyper-converged Virtual SAN and MaxParallel for Windows Server software across traditional storage, hyper-converged, cloud and hybrid environments.

DataCore's newest product, MaxParallel, launched October 2017 and works with the parallel processing technology that the company introduced in its flagship SANsymphony software-defined storage to improve storage performance. The MaxParallel workload optimization software is designed to remove I/O bottlenecks that can slow application performance.

Dave Zabrowski

In a recent interview, Zabrowski discussed broadening DataCore's product strategy, which includes overcoming skepticism around MaxParallel.

What interested you about DataCore?

Dave Zabrowski: DataCore, for me, was the perfect blend of an established, profitable, growing business that still had a lot of legs in its organic state coupled with the opportunity to take new technology into a brand new category that we call workload optimization. This new technology came out of some of the techniques that were being used to enhance storage performance. The DataCore brand MaxParallel got to market last fall, and we're in the process of taking that into a much broader set of applications.

Is DataCore trying to move beyond storage with MaxParallel?

Zabrowski: Yes, we absolutely are. If you uplevel what's going on in the industry and look at it from the customer down, what are the major trends? It's all about digital transformation. Every single CIO, regardless of vertical, the size of the company and geography, they all are undergoing their version of a digital transformation journey.

And if you look at the key elements of digital transformation, it's all about real-time data. And real-time data has to do with performance. It also has to do with the availability and the resiliency of that data. And that's essentially what storage does. If you look at DataCore's approach to storage, through our real-time synchronous mirroring, our replication of snapshots, the granular CDP [continuous data protection], it's all about making sure that data is always on, always available. That's real-time data. We then augmented that product suite with this new workload optimization MaxParallel product.

Is MaxParallel intended for use with DataCore storage or any vendor's storage?

Zabrowski: It could be absolutely any storage.

What type of applications is MaxParallel designed to accelerate?

Zabrowski: The applications could be standard third-party ISV-led applications -- things like Exchange or some sort of analytics platform or CRM. Or it can be for a homegrown application. There are a lot of opportunities for accelerating databases, which tend to have a lot of performance bottlenecks, and then even in the infrastructure -- not only for storage but also for virtual machines on the compute side, for networking and even for things like containers -- where MaxParallel can get by and accelerate it.

The way to think about it is anywhere that there's an I/O-constrained IT resource, MaxParallel can help eliminate or dramatically assist in eliminating those bottlenecks and improving the performance. Typically, we'll see somewhere between a 20% and 50% performance improvement by using MaxParallel.

How does MaxParallel work?

Zabrowski: For whatever I/O that's waiting to be processed in your cache, that's single threaded, MaxParallel takes it and spreads it over any number of cores. Let's say you're on a four-core system. You're I/O bound on your first core. MaxParallel will sense that and will take that queue -- that cache that's waiting to be processed on that one core -- and fan that out over the other three cores so you'd have more of a balanced process.

Where does the MaxParallel software run?

Zabrowski: It's nonintrusive. It does not disrupt the kernel of the OS. It does not go inside the application. It basically sits at the driver level between the OS and the IT resources that it's accelerating. The technology is very simple to deploy. You download it from our website. And all you have to do is restart your application, and you'll get the benefit of it.

Does MaxParallel only work with Windows Server?

Zabrowski: Today, that's the case. We'll be making an announcement about a Linux equivalent of that product.

Is MaxParallel any different than the parallel processing technology that DataCore has been talking about for years within its storage products?

Zabrowski: It's the same. It's taking those nuggets that are part of SANsymphony and breaking them off and cloning them into its own product. One of the secret sauces is the parallel I/O processing capability that makes SANsymphony run so fast. We just took those assets and essentially cleaved them out from the product and created this separate product called MaxParallel.

How has the reception been for MaxParallel so far?

Zabrowski: To be totally frank with you, skepticism, because people have never seen anything like this that is such low friction that doesn't intrude into the application or the kernel to provide this benefit. It doesn't work magic on 100% of applications. There are a lot of applications that aren't I/O bound. But for those applications that are I/O bound, it's literally like magic.

I was born in the state of Missouri, and Missouri is called the 'show me' state, and that's what the customers are saying: 'Well, just show me.' It sounds too good to be true.

Do you see one of your primary objectives as expanding DataCore's mission beyond storage?

Zabrowski: I do. The storage market has been a very good market for DataCore. But I also see the opportunity to uplevel DataCore's relevance throughout the entire software-defined data center -- not just storage.

Do you see much opportunity for innovation in storage at this point?

Zabrowski: If you look at our synchronous three-way mirroring that we released in the fall, that's pretty innovative. With our CDP capabilities, you have the ability to essentially do auto backup and snapshot backwards, retrospectively. We've seen that innovation adopted in areas particularly where data availability is mission-critical.

Other things that we'll be releasing later this year [involve] some hybrid cloud work. We're bringing our three-way mirroring to a hybrid cloud where you can seamlessly take your on-prem storage capabilities and mirror and migrate that to a hybrid cloud model. We're also working on Linux versions and continuing to add to our high availability and disaster recovery capabilities.

What's the impetus for adding Linux support after all these years?

Zabrowski: Linux is a good chunk of the market, and certainly as you go into larger enterprises above midmarket, it's table-stakes. And we're going to pursue that so we can participate and serve those customers with the same innovations that we have [provided] the Windows-based customers. It's just basically we're broadening our appeal.