Monday 19 January 2009

Network Analysis

At the moment I am researching which is the best tool to buy for an enterprise network in terms of discovering network performance and so I thought I would ask any readers of this blog for their recommendations.

First a bit of background I work in a support team in a large enterprise and we often get calls reporting the "network is slow". Now we have some tools that help us monitor the utilization of the links but they cannot show us the top talkers across the links. Our current tool set also does not allow us to see what protocols are being used across the link and we have no chance of being able to prove what application is causing the problem. It always turns out to be an application or patch deployment that causes the problem, however we usually have to let the users suffer from a "slow network" until the problem goes away. To me this is unacceptable service and I have started investigating products that can give us a bit more visibility into the network.

Ideally what I would like is an application that can pull NetFlow stats from the routers and via some sort of probe be able to investigate switched traffic. I would require the product to determine application responsiveness and ideally be able to build a baseline over time. We may also be consolidating data centres fairly soon and so discovering the relationships between servers would be a great bonus. As I want to make sure we migrate all the servers that talk heavily to each other as a package and we do not leave a server behind that then hammers the WAN link by talking to it's peers in the new data centre. Finally I would be interested in how the various products cope when WAN optimization is introduced into the network.

I have been looking at tools from MAZU, OpNet, NetScout and NetQoS but if anyone else has any recommendations I would be most grateful.

Wednesday 7 January 2009

State of the Ferret 07/01/09

Well a belated Happy New Year to everyone. It has been a while since I have posted and that is because things have been pretty hectic with family commitments over the festive season as you can probably imagine. There were a few work commitments in there as well but studying did not go very well at all. I was trying to finish the OSPF sections in Narbik's workbook before Christmas, but I have only just got round to doing it now.

There comes a time in life where you have to face up to reality. My reality is that I am not even going to get close to being ready to do my lab in April. I have a second child on the way soon and I am only just beginning to remember how much work a new born child is. This time it will be especially challenging now that I have a toddler as well. With my wife going off on maternity leave imminently money is going to be tight and any free time I get will have to be spent focusing on doing overtime (providing it is not stopped) to make up for the shortfall in our finances. It will probably also pay to keep focused on the job a lot more in the current market, to try and ensure I do not join the massed ranks of the unemployed. So in short study time is going to be very scarce or non existent. This in turn has hit my motivation which is leading me to think I will never get round to doing the CCIE lab exam. I am trying to retain some of my knowledge by doing mini labs whenever I can but for now it looks like there will be no full scale practice labs for 6 to 8 months at least ouch :-(

On the good news front I did manage to get into the Everything IE beta but as it is confidential I cannot say anything about it at the moment.