From time to time I happen to participate in social network discussions among IT professionals about cloud computing. Lately, the main topic seems to never stay focussed on the cloud itself but on the aversion the operators (system integrators, VARs, resellers, etc.) have to this new approach to managing enterprise IT.
Such fear is often due to considerations about cloud computing, specially public clouds, will more and more erode sales of companies that sell technology infrastructures with a consequent loss of jobs. I’ve also lately seen some players to openly oppose the phenomenon just finding it hardly credible, without stating any technical reasons, just considering cloud computing an evil occurrence to be carefully avoided, spreading FUD on their customers fearing the possible loss of a job! The argument often sums up: “If you adopt cloud technologies and simplify things too much you’ll lose your job because you will no longer be indispensable!”
I can’t help but strongly oppose this message, I think that things really are in a very different way, here is why:
not all clouds are alike!
You can find many different things sold as cloud computing, often falling into ridicule (Telecom Italia, the most important italian telco operator, even set an urban lighting project as cloud computing).
Anyway, to sort things out, cloud computing is simply a model where you pay just for the resources you actually use, not a fixed amount beforehand. Cloud computing is the outcome of ASP, then *AAS, evolution. Technologies involved (web, bandwidth, virtualized resource management, security, applications) are sufficiently mature now to guarantee a confident approach to cloud. Cloud is just an evolution from these technologies, it is a paradigm we will have to second, withstanding its siege might just slow it down a little, but as soon as the dam will break, the outcomes might be devastating.
If I were an IT manager, I’d put cloud computing under close scrutiny and, if possible, would test it with my own hands. Being able to buy a service billed according clear rules that take into account its real usage (number of users, CPU cycles, rows in a DB) is really inspiring, it would give me full control of expenses for present use and guidance to what to budget for according usage trends. Some applications can be easily find their home on the cloud (as a rule of thumb, most of those born in the internet era), while some simply cannot fit the cloud model (just a thought about mainframe technology, right from the ’60s and and more or less widely still present). Simply put, a software to be used as SAAS must be designed from the start keeping well in mind its cloud character, it will be otherwise very difficult, to say the least, to implement and use it on the cloud.
As usual virtue stands in the middle. IT growth is almost exponential, an IT manager is often asked to set up more and more services, some to be used by the company, some more to be used by outside (customers, suppliers, partners).
The amount of data to be managed keeps on rising at a very fast pace, companies struggle to keep expenses under control.
Cloud computing can be a way to rationalize budgets and save money. A radical approach, as I’ve learnt in more than 20 years of experience in this field, usually does not pay. Going totally public (i.e.: amazon, google, salesforce, etc…) is almost impossible: there are always too many links with legacy processes, old applications, bandwidth and latency issues and so on. On the other hand, you cannot stand still, hoping that nothing in your company will have to be changed.
Current infrastructures must evolve otherwise will soon become unmanageable, automated processes will be increasingly necessary, provisioning processes will have to be revised and simplified. This will lead to a private cloud (at least infrastructure or platform: IAAS, PAAS).
Hybrid cloud will win, I keep on repeating this since several time. You will not be able to keep everything inside your company walls, and you will not be able to migrate every IT aspect on a public cloud. A sensible strategy will be to use both public and private cloud infrastructures to maximize the result.
In the last year I’ve seen an interesting project: a customer coming from a complete outsourcing strategy decided to go hybrid, with a SAAS ERP (SAP) hosted by a market leader service provider, and has built from ground up an in-house virtualization infrastructure that will host every other IT service. This customer is already able to measure the savings made, at the same time preserving the labor force and improving the quality of service and response times.
Surely it is a long road to go, no standards are defined yet, there are no tools to interconnect public and private clouds, and expertise is still to be trained, but be sure, the whole IT industry is taking it.