When I think about clouds, visibility is not something that comes to mind. Taking the complexity associated with today's service infrastructures and putting all (or parts) of that into a 'cloud' sounds more like flying on instruments to me. Maybe that's why many customers are looking for dashboards….
At least with instrument flying there are flight rules that establish universal procedures and regulations for ensuring that we don't crash; I'm not aware of any such thing for cloud computing.
I like a recent post by eG Innovations that references an article that essentially referred to cloud computing as an operations model rather than a technology. An e-book I found referred to cloud computing as a model of service delivery resulting from the convergence of:
• Hardware standardization
• Virtualization and other software advances
• IT management and business practices
I know one thing though, when I hear a Ph.D and former research scientist say,
"All I could get was “cloud proven virtualization and networking products”, “most widely-adopted virtual infrastructure platform for hosted cloud services”, “cloud-proven virtualization platform”, “cloud-scale distributed virtual switch”, “cloud balancing”, and “burstable clouds”. It all went over my head!",
I know the spin cycle's in high gear.
While there's no question that clouds promise much, I'd like to remind us that we have been here before. Telcos have been 'clouds' for some time now... how much do you trust your friendly Telco provider?
Here's a small example of what can happen when we try and fly on instruments without the proper regulations, even when the services provided are actually pretty good. (It was, as is often the case, mostly pilot error)
The commenting on this weblog was provided by Haloscan, who was acquired by JS-Kit. This happened so seamlessly that I was totally unaware of that, since I do not recall being sent any message announcing that fact. Perhaps because my web builder software (Sandvox) had links to Haloscan built into the software. However, even the JS-Kit site states,
"Haloscan, the legacy comment system that JS-Kit acquired last year, is physically starting to fail (the software and hardware). In order to minimize the disruption for users and avoid a hard stop, we have worked hard to provide two ways to transition off the system."
Even when I upgraded to Echo (the new offering from Haloscan, JS-Kit, whoever…) I had some issues associated with moving ov
er admin rights, commenting, etc. What was even more interesting was that support was provided by GetSatisfaction (yes, another cloud) who seemed to enjoy asking "how this made me feel"...
You get the rest of the picture…
This is a very tiny example of the kind of chaos that 'cloud' computing can wreak if we're not careful, and brings back bad visions of telling customers it was 'the local loop'… if you think that the service demarcation points between applications and infrastructure is going to be any clearer than it was between local loops and frame relay clouds, I've got a bridge to sell you.
Security and privacy concerns are not the only things we should be worried about. Cloud management has the potential to be the real nightmare, and the cloud wars may obscure this fact just like clouds abstract IT complexity.
Cloud transparency from a monitoring perspective would provide all stakeholders involved in an 'end-to-end' service infrastructure with the ability to see what's happening at every layer of every component in every cloud -- network, system and application -- and automatically isolate which layer of which component (from which cloud) is the source of an anomaly. The only dashboard that matters is the one that can keep you from flying into a cliff...
I have no illusions that this is going to happen anytime soon. More than likely we'll get competing management schemes that place the burden right back on the customers, subscription plans that are almost impossible to terminate and continually renew,
and upgrades to the next generation when the cloud is ready (not when you are). Yes, the road to IT service management excellence is taking
yet another very savage turn...
I know of many customers who have no business processes defined (and certainly not from the Outside-In) and have virtually no services defined either (no virtual pun intended). That means no SLAs mapped to business requirements, no exit strategy or end-of-life plan, yet are excited about the prospects of moving IT complexity to a cloud (or multiple clouds) ...
Hmmm, a region of the service provider space where nothing can escape; something undetectable that marks the point of no return....
That's not a cloud, that's a black hole.