The illustrious Stuie blogs about latest Cloud outage here and opines about the new paradigm and how application developers need to change as well. I think it’s something worth examining, infrastructure guys are getting a kicking from all sides about how they need to change; dev-ops wants to come along and ‘steal our jobs’, users don’t believe we are responsive enough, we don’t embrace change and so the lists go on.
But surely cloud needs a change from everyone and possibly a step back to ensure what we are calling Cloud is not simply a slightly more responsive and agile deployment of infrastructure enabling applications to be rolled out slightly faster but with little underlying change. And I am not just talking about the Cloud-wash of server virtualisation.
Applications themselves need to become more resilient in themselves; they need to become resource aware and become able to request more infrastructure resource but also release the resource when not needed. They may even need to become aware of their own environment and be able to query a centralised resource broker which will move them to an appropriate resource; an application may become aware that it has become docile and it can be moved to a state resembling hibernation. It may become aware that it is handling more transactions and ask for more resource or even to move to a larger resource; it could even be aware that it has a single transaction which has caused a temporary spike in activity, a badly written query for example and it asks the broker whether it is possible to have temporary capacity on the infrastructure it is on.
We may even be looking a new operating systems which are much better in handling distributed processing and instances.
And then there is the role of the user and how do their expectations change. I am not convinced that some of our current user experience metrics stand-up long term and in Enterprise IT, have we fully taken notice of what else is happening in the world around us? We have some odd dichotomies; our users seem to expect instant gratification except when they don’t.
I don’t know if you have watched your children online; it’s kind of odd, they expect excellent performance but tolerate shoddy performance. If YouTube runs slowly, they are not exactly happy but they will find something else to do instead; they will come back later and try again. Of course, if the whole environment runs slowly, then you’ve got trouble but we have a generation of users who tolerate inconsistency in performance a whole lot better than their elders.
But we ourselves are changing as well, if Amazon is down; well, we’ll probably wait to order that book. Yes, there are other places we could buy that book but we like the Amazon experience and most of the time, the service is good.
I think that as we move more applications into the cloud; this experience will become common. But business processes might have to change to ensure that a user who is bottle-necked in one application is still able to progress in another application. Workflows will need to be looked at in that mind.
Can we define roles with this flexibility, is it even desirable? Is it even inevitable as attention spans and a youth spent in distraction changes us?
You see when people talk about Cloud being more than just technology, I’m not sure whether they realise quite how right they may be.
Martin
Some excellent points I’d like to reinforce here. First, on watching how your kids use IT — and how they use information. Best way ever to get insight on the future of IT.
My generation believed that information was static, precise and lived in a handful of known locations. This generation believes that information is volatile, imprecise and can live anywhere.
My older kids are far better at assembling imprecise content to create unique insight and perspective than I ever was. Me? I was taught to grind spreadsheets and make pie charts 🙁
Look how quickly they’ve ditched the WIMP (window, icon, mouse, pointing device) paradigm in favor of finger-friendly tablet applications. A far cry from the command lines I grew up with.
Enough on that, back to cloud!
Yes, multiple aspects need to change — no doubt about that — the challenge is “in what order?”. Many people have decided to attack the infrastructure problem first, in the light of having to run legacy apps that are unlikely to be rewritten anytime soon.
— Chuck
Attacking infrastructure is probably the easiest and takes the least work but the risk is that actually nothing changes. Or what happens is that green-field developments leapfrog you and their agility and ability to leverage Cloud kills you. I suspect there are a number of companies already who are dead-men walking and these are not IT companies. You continue to paper over legacy and give it a veneer of Cloud at your risk.