This week I have to host a panel debate on “stress testing cloud applications and infrastructure” at Netevents in Rome (I know – it’s tough, but someone’s got to do it…).
One of the areas to cover is, well, how you do actually cover that kind of environment from a test perspective – e.g. engage thousands of what we used to call human beings to all use specific apps at certain times, or can we simulate that or… given that we live in a world of analytics – well, we always have done, just that now they are being collected and – funnily enough – analysed, is a lot of the hard work actually being done for us? I mentioned in my last blog that I recently met up with John Rakowski of AppDynamics, the Application Intelligence company in the Enterprise space (that’s as in the type of business, not the starship – well, not yet at least) and a couple of areas we talked about were application intelligence and unified monitoring. In other words, the ability to blanket monitor, so you are collecting all the data into a unified reporting mechanism, and the ability to understand the app’s you’re actually monitoring.
This is a gazillion miles away from the old methods of collecting and then sifting through Syslog files and other Data Centre consuming information logs, requiring several of those human being things again on hand to manual carry out this most exciting of tasks – finding the eNeedle in the data haystack. So, in one fell swoop you minimise costs, remove human error, and maximise visibility and the ability to pro-actively manage apps and services.
It’s much the same kind of story in the world of network monitoring itself; I had a catch-up last year Savvius (the artist formerly known as WildPackets) and gone are the days when we had to search manually through disks-worth of Hex in order to find a particular packet identifier or character string for example. In its recent update, you can now correlate and analyse network data directly on the capture engine, as it happens, and it give you remediation advice too!
So, back to the original point – are these, let’s call them “app and data visibility tools” actually doing the job of a specialist app/service product testing, er, product and, more worryingly, that of the product tester?
That’ll be one to debate in Rome then! Bring on the pizza and Chianti (classico reserva, of course..) .