EMC has made a lot of acquisitions over the last few years.
Some get a lot of visibility -- VMware, Documentum, RSA, etc. -- but there have been many others that have turned out to be shining success stories in their own right - just a bit overshadowed by the big names.
A lot of people think that acquiring companies is hard work. They're right.
But exploiting the full value of what you've acquired -- that's really hard work. And there's a general nay-sayer crowd that handicaps each and every acquisition EMC makes, including today's.
Yes, we've had a few bumps, but -- in general -- I think EMC has done an exemplary job of making acquisitions that deliver on Joe Tucci's famous "two-fer" comment: look for acquisitions that get us into interesting new parts of the market, and at the same time make the core stronger.
As one example, let me share with you the story of Kashya (now RecoverPoint) as an example of what happens behind the scenes when good technology ends up in good hands.
A Bit Of Context
Most everyone knows that EMC has built a pretty substantial franchise around remote replication.
SRDF (later followed by MirrorView, Celerra Replicator and a whole lot of other products) became the de-facto standard in how large organizations kept a synchronized remote copy of their data available for recovery should it be needed.
Compared to other approaches, the idea of replicating from an array turned out to be much more successful than replicating from individual servers, which was the other option at the time.
But a few years back, we could see that this market was going to change.
Change Was In The Wind
The first trend we saw was that there was a lot of talk around a new form of replication -- CDP, or continuous data protection. Kind of like Tivo for your data center.
The replication model up to this time was either synchronous (exact real-time copies at both ends), asynchronous (remote end lagged by seconds to minutes), and point-in-time (make a clean remote copy periodically).
But there was interest in being able to rewind the action to any arbitrary point in time. Let's say you had a bunch of databases teaming up to run an business application. One of them got corrupted, and then the corruption spread.
Wouldn't it be great to rewind the movie to the point just before bad stuff happened, figure out what happened, and play it forward from there? And do it while protection continued?
That's the basic idea bhind CDP -- a bit oversimplified. If EMC was going to grow with the market, we'd have to add this capability at some point.
But there was a second, more disruptive change.
More and more people wanted to do replication in the network, rather than inside the array. They wanted the freedom to make independent decisions for storage and remote replication, rather than one decision for both.
And, if a protected application happened to touch multiple arrays, it's easier to provide consistency using a network approach rather than coordinating multiple arrays and multiple hosts.
Now, in certain corners of EMC, this was a bit threatening. We had this very nice franchise around remote replication at the storage level, and the idea of moving it outside the array was, well, a bit heretical to some.
People could get remote protection without having to buy EMC array products. What would happen?
To complicate matters further, a few vendors had tried to offer replication appliances that lived in the network, and none of them had done too well. I seem to remember HP getting a bit bloodied with this idea, but they weren't the only ones.
But most of us who aren't tied to a single product family realized that we could disrupt, or be disrupted, the choice was ours.
Our First Take
Interestingly enough, the impetus for addressing this requirement came from our Legato Networker group. They had been looking at this problem for a while, done some architectural work, and saw this more as a software problem than a storage problem.
Incidently, one of the nice things about making acquisitions is that you get really smart people who look at the world differently, and may see the upside of new things rather than the downside. Put differently, it's hard to imagine a Symmetrix product manager sponsoring the idea to put replication in the network rather than the array.
We got to market pretty quickly by working with Mendocino software, who had pieces of the technology that we needed. The good news is that customers were extremely interested in the whole idea of CDP, and similarly interested in the idea of putting the functionality in the network, rather than in the array.
But as we got into it, we realized that we needed a richer technology base, and that this was going to be a pretty big deal.
And Then There Was Kashya
We had run into Kashya in a few engagements. I even had met a few customers who we thought were really, really smart, and they had started to ask us some very pointed questions about Kashya. These discussions didn't trigger a defensive reaction as they do at most large NIH (Not Invented Here) companies; we were intellectually curious about what they'd done, and what made it so appealing.
We ended up being really impressed by the technology and the people. We liked the architecture, the user interface, and we felt they'd figured out a few very subtle yet important things we thought showed their understanding of the space.
But, like most small companies with good technology, they had a daunting task in front of them.
In the IT and storage world, remote replication is like open heart surgery. The stakes are high, it's complex, support requirements are stringent, and it's not something you're likely to trust to a small, unknown company. They were winning a few deals, mostly in smaller companies, and it was turning into a long hard slog for them.
And, from a business model perspective, they looked like they could be in a gravity well -- they needed more people to drive revenues, but that takes money and time. Ultimately, they wanted to see their technology succeed in the marketplace, which is what EMC wanted as well.
And so the deal was done -- and then the real work began.
The Deal Gets Done
When we announced the transaction, there was a bit of friction within EMC. Kashya was obviously an entirely new take on remote replication as compared to our array-based approach.
What do we tell our customers? What do we tell the industry? We didn't want to oversell our new capabilities, but we thought them important enough to say something.
And then we started to find more and more opportunities for synergy ...
Finding Synergy
One thing that challenges most smaller companies offering remote replication products is the extremely daunting qualification and support requirements that reflect the diversity of the environments customers actually use. And, trust me, customers won't rip and replace their existing infrastructure just to use your shiny new toy, no matter how cool it is.
We're talking lots of different storage arrays: EMC, HP, Sun, IBM, NetApp et. al.
We're talking lots of different switch and HBA configurations.
Lots of different servers and operating systems. And, of course, a wide and diverse application landscape: files, Exchange, Oracle, SAP and more.
Well, EMC had done that sort of work for more than a decade, and had built up an enormous amount of investment, expertise and process, mostly reflected in EMC's eLab. As of the day of the acquisition, this billion-dollar-plus capability was immediately available to the new RecoverPoint (formerly Kashya) team. Hard to see how a smaller company would be able to match this.
Another requirement is customer support. Remote replication environments are very complex and have a lot of moving pieces: networks, replication engines, servers, applications, storage, etc. As some people have found, they're not nice and tidy environments to support at enterprise-class levels.
And, in the big landscape of things IT, remote replication is pretty sensitive stuff in terms of attention it gets when it isn't working as adverstised.
Once again, EMC had done that sort of work for more than a decade. We had literally thousands of support people who knew remote replication and had been exposed to tens of thousands of customer environments. We had the expertise and the process in place day 1 that could be brought near-immediately to the RecoverPoint solution.
Again, hard to see how a smaller, aggressive start-up could match EMC's investment in this regard. And, even if you find people who take your product for free (as many smaller companies are forced to do), it won't be high-profile implementations that you'll get.
Not every customer is an expert at defining their requirements, choosing the right technology, implementing it correctly, and learning how to manage it day-to-day. Once again, EMC had a robust consulting capability built out around our existing technologies and customer set, so this was pretty much a drop-in for RecoverPoint as well. Can't see a small startup building out a robust professional services capability around a single product.
There were some other interesting technology tie-ins as well.
EMC's Replication Manager could easily be extended to help customers manage RecoverPoint replicas the same way they manage TimeFinder, SRDF and other forms of replication. Nice synergy. And EMC had a pretty robust set of CLIs and APIs customers used for scripting automated replication tasks -- again, a nice form of synergy.
On the application front, we knew already how to work with Oracle RMAN to get hot backups, and Microsoft's VSS to get consistent splits -- now part of the RecoverPoint offering. All already part of our capabilities, we just brought what we already knew to RecoverPoint.
On the architectural front, we believed that intelligent switches from Cisco/Brocade were pretty important here (as opposed to being limited to just appliances) -- and, of course, the work we had done with Invista previously was a big help. Today, we've got a nice portfolio of happy customers running CDP in production on intelligent switches, and unfortunately the smaller guys have to say "roadmap".
Finally, EMC was already having hundreds of discussions each and every day regarding the whole topic of business continuity.
We got comfortable rather quickly about positioning where RecoverPoint fits as compared to SRDF, MirrorView and the whole suite of remote replication products EMC was already selling.
We didn't have a situation where we had to force-fit a single product into every customer environment. We could step back and have an honest discussion around the pros and cons of different approaches.
Again, this was a luxury that smaller, single-product companies can't do for their customers.
And It Started To Work
Within a year of the acquisition, not only had we been able to go back to the initial Kashya customers and tell them what we were doing with the product (all good, by the way), we found that more and more EMC customers who were already using our existing remote replication products wanted to put RecoverPoint on the floor and put it through its paces.
We've exceeded our own (rather aggressive) targets in terms of customer adoption. And that at EMC becomes a virtuous circle -- the more customer interest, the more R+D, the more sales and marketing attention it gets, and so on.
We also get an opportunity to link what we're doing with something like RecoverPoint into other hot areas of our portfolio, like VMware. Turns out that RecoverPoint is a great strategic fit for people building out new VMware farms, and want the ability to granularly replicate and recover individual VMs, and still keep the infrastructure external to the server.
Lessons Learned
Sometimes I think it's a bit challenging for all those promising tech companies out there, especially those that compete with areas that EMC has invested in. It's one thing when it's a level playing field, it's quite another when a big company like EMC moves in, picks its favorite, invests heavily, and creates a very one-sided situation.
All that's left for the other guys to go do is hope that one of EMC's competitors feels threatened enough to make a responsive acquisition (something we've seen more than few times) and make a go of it. But it's never as good for the after-the-fact competitor, or the acquired company, as compared to the first movers.
Without going into too much detail, I've seen our larger competitors struggle more than a bit with fully exploiting their acquisitions. We struggle as well, but -- fortunately -- I think we're doing far better than most.
And this story has repeated itself (with a few variations) for Rainfinity, Smarts, Avamar and other parts of our expanded portfolio.
I think the trick has been Joe Tucci's philosophy -- look for two-fers that get you into exciting new parts of the market, yet can leverage and strengthen the core.
Comments