Pro Curves

This last Tuesday I took some time out to head over to a HP Procurve Seminar that was going on across the street from my office. We’re moving to downtown KC this fall and we’ll be putting in a new network infrastructure and most likely a new Voip system. The Procurve switches are on our short list of equipment possibilities so I thought it would be good to go taste their brand of Kool-Aid. Currently we run a little Cisco and some Extreme so this would be a new venture.

I would say that the better part of the presentation was when John McHugh got up and spoke about the Procurve vision. I liked the idea that they weren’t looking to branch out into other things, really their plan was to make switches and make damn good ones. Switching isn’t a business that is going away any time soon having a company focused on working that market is a great idea. I think Cisco has lost some of it’s way in that area as they have branched out, and in trying to do so many things for so many people have lost the focus that’s needed to keep a pinnacle product. Interestingly though Mr. McHugh spoke a lot about the switch becoming a sort of virtualization backplane. I’m not a switch guru or anything but I have noticed more services moving out to the edge, your department switches, when previously almost all of that logic happened at the core, your big primary switch. If I understood correctly some of this virtualization idea is to run things such as firewall, virus protection and things I would term network stream services within the switching architecture rather than outside appliances. As HP doesn’t want to create these services they would be looking for partners to work with them, first creating blades which fit into the switches that perform these functions and then possibly moving them to virtualized packages. From the vendors I talked to recently I have the distinct impression you won’t be getting A-list vendors to do this, but would probably generate a whole lot of interest in B-list solutions looking for an edge to move up. This take on what the plan to do makes the Procurve vs Cisco models eerily familiar to the Redhat vs Microsoft models. I think virtualization of services and not only serving these needs at the core but at the edge isn’t something that people will jump to adopt at the switch level any time soon, but it’s time will come around and Procurve is getting to the point where they are starting to lead and they can’t afford to give that up.

There was a talk by Abner Germanow from IDG, from which I pulled that companies in these tough economics times are pulling away the least from networking projects. I believe this mainly has to do with companies coming into Voip and the utter explosion of fat traffic on the internet the past few years. His point was that it was mainly due to video but I think the availability of broadband to the home has also ratcheted up what kind of heavy content companies are willing to spit out to the client and what people pull in at their office.

Barbra Stuter gave a live demo of Procurve switches and showed us a neat marketing video of how they can blow up a server room and have failover happen quickly. The demos were spiffy but I’ve come to expect fail over, so seeing that it works wasn’t anything all that great. It’s kind of like someone demonstrating how quietly a new car starts up, at this it’s not a big whoop because you just expect a car to start up now. In fact I was actually kind of shocked that in their video demo it took two minutes for the Windows and Linux servers to failover, that seemed a bit long to me. She was a good speaker though and I would say of all the presenters she was probably the best communicator. One of the things she talked about was automating some of the network management decisions and how it might be a bit scary to give up some of those decisions. I’ll just say that right now I’m all about automating pieces, I’m fast approaching 100 servers and there is no way in hell I can hands on manage everything that needs to be done. Bring on the tools so I can twist the knobs to get it to do what I want leave it alone,  and later report back to me in a good summary report. There was a lot of talk about open-ness and I need to go look and see if the APIs HPs management gear are using are open, it seems to me they are doing more than what you can simply do with SNMP.

George Hleback from SonicWALL and Joshua Haslett from Mitel were both there pitching their products and gave good presentations. Mr. Hleback did a pretty good job of putting the fear of God in me, he made me happy I manage our facility and don’t have to deal with gaping security hole that can be desktops and users. Mr. Haslett’s talk on Mitel was good, as I haven’t been the one doing the deep research into the phone systems for our move he managed to leave that damn sales nugget in my head that makes you wonder if maybe their product is the one to go with. The wireless phone he had with him that also had the ability to VPN tunnel back to office and display html within the phone display was definitely “the shit”.

It was a good seminar and I’m glad I went, I kind of wish I had stuck around from lunch and could have done some more Q&A with Mr. McHugh. Tuesday’s though Kate comes into the office, it wasn’t a big enough thing for me to miss out on my once a week lunch date with my wife.