Tuesday 7 August 2007

F5 buys Acopia

Here's something interesting. I've worked with F5 in the past, I've trained on their kit to advanced level, I also have (fairly infrequent) contact with Lori MacVittie of the SBN, who is a very helpful lady and always cheerful. For those that don't know the BigIP, or knew the BigIP on version 4, before version 9 was released, do everything you can to get a look at one now. They are quite simply fantastic.

For anything load-balancing, this should be your first and last stop, you'll never need anything else again. They have also made a number of acquisitions in the last few years, SwanLabs for the WAN acceleration, Firepass for the SSL VPN box, and I can't remember the name of the TrafficShield manufacturer, but that one too. They make Link Controllers, DNS management systems, load balancers, and all of the above.

Their most recent foray into the application firewall world was something had my doubts about, but they really seem to understand it. Along with the incredibly powerful iRules and iControl language for controlling the input and output of the boxes, they are one of the most usable pieces of kit I've ever laid my hands on. Granted they are expensive, but when you are a distributor, that doesn't really concern you so much.

So why have they decided to buy Acopia? First of all, this article is a bit short-sighted. Integration with SwanLabs has taken a while, but they've done it well, and better than other players in this space, who they were racing against. In the end, slowly and surely will win the race. Acopia will fit nicely with their currently portfolio. Something I often wondered about F5 was if they would get into controlling storage, or just content, and then how. It seems they've been thinking about this themselves.

Apparently they are going to make their file optimization part of Acopia's WAN acceleration product to start with. This means integration with their existing offering, WANJet, which already does this, so they obviously have a plan there too. Presumably they have found a way to integrate the whole suite more easily if they integrate with the Acopia software first.

Anyway, that's not really the point, what is the point is that having already got a really good control of connections into the network, they will now have control of the network as it moves into storage. This is a great move for them. Notice they aren't mucking about with user security (yet), why reinvent the wheel when you can build up a backend which rocks, and then integrate it with some other killer auth system, of which there are a multitude.

This is why I've always liked F5, they started in a space which no-one else really had much of a clue about, load-balancing, and did it really well. They are better than Cisco's attempts in this area, easier to use than Foundry, and others are just toys. But they didn't stop there, they've tightened up the security on the way in, splitting networks between the inside and the outside. Now they have WAN acceleration, which ties in with load balancing far better than firewalls or proxies (Juniper and Bluecoat also brought out WAN accelerators at around the same time).

They seem to be in the right place in the network to do a lot of things. Before Hoff moved from Crossbeam we had a number of exchanges about Data Security. My conclusions from them were that we would soon be seeing convergence with a box at the perimeter to control users and apply security of all types on entry into the network, and one nearer to the storage to control security of the data. I think this is probably what F5 are betting on too.

I hope they're right, it would be a great win for data-centric security.

No comments:

MadKasting