Riverbed goes with the flow

Its successful IPO hit the headlines, but it wasn't all plain sailing to get there, says Riverbed VP Alan Saldich.


Two months ago, Riverbed Technology became the first big IPO in the storage and WAN sectors in years, when it raised $86 million in a public offering of shares - shares which were trading at a premium of 60 percent on the offer price within just a few hours.

"The day we went public was very exciting," says Alan Saldich, Riverbed's product marketing VP, "but on the other hand everyone had work to do. So there was a glass of champagne and then back to work."

Saldich argues that far from being a sign that the dotcom madness of the 1990s could be returning, the success of the Riverbed IPO - which has given a fillip to the whole WAN optimisation industry - was simply the pay-off for the four-year-old company's hard work.

"We're not profitable but we're very close," he adds. "If you back out the stock options, we only lost $1 million last quarter on revenue of $25 million.

The odd thing is that, when Riverbed's founders were looking for 'the next big thing', WAN optimisation wasn't on their minds. They thought it would be webcaching - CTO Steve McCanne had already developed the software - but it all changed when they did what many others forget to do - they asked potential customers.

The results, says Saldich, showed that no-one was was interested in buying webcaching, not least because open source tools such as Squid did the job and were free.

"But enterprises said they had other problems - file sharing, remote email, portals... We couldn't fix those at the time but Steve got to thinking and came up with a general solution," he says. "First, it should not be a cache; second, you've got to have disk to store lots of data; and third, be a TCP proxy, not as server proxy."

More than just WAFS

Saldich is adamant that, although Riverbed won many of its early customers on its ability to accelerate access to centrally-stored files over a WAN, it never aimed to be just a WAFS provider.

"We were never convinced that it's just about files - Cisco was though, so it bought Actona," he says. "Our approach was to do things in a way that would work for any application, for example remove repetitive data and make TCP more efficient, but then you need to do something for applications."

Some have argued that the market for WAN optimisation boxes is a transitory one, and that the technologies involved will eventually migrate into the networking layer, becoming a standard component of branch office routers for instance.

Although that is already happening to some extent - for example, Cisco has application acceleration cards for routers, Blue Coat has added WAN acceleration to its gateways, and Juniper has said it will follow suit - Saldich claims that they won't be able to do all of what Riverbed does. He says that, just as it's not only about files, it's not only about networking either.

"We don't think it's possible or practical to put what we do in a router," he asserts. "Maybe you could with what Cisco and Juniper do - they are primarily reducing network bytes, and do very little at the application layer.

"We don't think this is a networking technology, because although networking is an important part of it, it's not central," he adds.

Ways of working

"What Riverbed does for a company is change the way it works. For example, there can be more inter-office collaboration, it makes projects less geographically-dependent. You no longer need distributed IT support - that saves millions of dollars. The impact is on the IT infrastructure overall, not specifically the network."

Saldich agrees though that - for now at least - there is room in the market for Riverbed and all its competitors. Yes, they find themselves bidding against each other to win deals, but their biggest competitor is lack of awareness, and the custom they're fighting over is a tiny slice of the total possible market.

"The active market is small, but that's how things start," he says. "Customers typically buy two, four or five boxes, then come back the following quarter and buy 20 or 30.

"For the likes of IBM, this market is infinitesimal - and a lot of people don't believe it works. It's one thing to find out about it, it's another to believe in it."

But he adds that the market is growing up fast, and there is much more awareness than four years ago - in part thanks to Cisco's aggressive push into application networking.

And he says that as well as trying to win new customers, Riverbed wants to push its technology deeper into the organisations that have already adopted it. It is therefore joining rivals such as Orbital Data (now owned by Citrix), Blue Coat and others in developing WAN optimisation software for individual PCs.

"We have over 1300 customers now," Saldich says, "and all those have significant laptop populations that want a software client."

So perhaps all those investors aren't so crazy after all.

"Recommended For You"

Blue Coat: Creating an economic advantage for users in 2009? WAN optimisation or a 'real' upgrade?