I do, however, spend most of my time talking to the most important people in our organisation: customers. Listening to them, and looking at the market, the way I see it is this: we are going to need many of the services of the ADC for the foreseeable future. Things like TCP optimisation, application firewalling, load balancing, event based traffic management, server offload etc. are not going to be offered by operating systems or client-server applications. These services are not networking, not applications they are application services delivered in the network.
What I think will change is what device the application services are run from. Today we have ADC's that are commodity servers running software (e.g. HAProxy), dedicated hardware appliances (e.g. BIG-IP) or virtual editions (e.g.BIG-IP Virtual Edition) . In the future I think that we will see ecosystems where the ADC components might run on a number of devices to create the required services. I see hybrid hardware/virtual systems for agile private clouds, with hardware devices doing the heavy lifting of optimisation and offload, but virtual editions of the same product providing discrete application traffic management. This will allow an entire application, or even an entire customer virtual datacentre, to migrate around an organisation's datacentres dependant on available resource, or to burst into commodity CPU to meet a peak workload.
The emergence of dedicated network devices designed for virtualised compute environments (using the yet to be standardised 802.1qbg/h standards) gives another potential 'home' for these application services.
The way we are going with Traffic Groups and ScaleN in V11 of BIG-IP maybe gives an insight into the future.
Robert
