It’s something like one month I haven’t posted on meedabyte.This doesn’t mean that nothing deserving to be on the blog was happening :). Actually I’ve been quite absorbed from writing an interesting (fully open source powered) architecture for a Business Process Management and Business Activity Monitoring facility for a client. This activity and others eventually made some thoughts come up to my mind.
I’m starting to think that we need new tools, both theoretical and practical, to manage the process of develop “solutions” at today. I use the word solution, and not something like architecture or software, infrastructure or service. This because, from an ISV, SI, or consulting point of view., is not more a software or a service that customers are demanding. Customers demands solutions for their problems:
- that can deliver value in short terms
- needing very low upfront costs
- able to scale and adapt to hopefully growing businesses
First of all we all need to change our mindset about requirements. With less to spend you need to focus on your real requirements. From one hand it means that customers are no more willing to pay for requirements that are shared globally since such requirements, almost surely, have generated in time one or more open source communities providing an open source solution. This trend is really clear if we think about software industry: former market leaders are loosing market shares because of open source newcomers. Look to mobile OS market with Android and Symbian, or to notebooks with linux. Most often, as the last effect, the original product gets commoditized.
From the other hand, the need for less up-front costs will unavoidably lead to more exploitation of cloud services. I’ve recently found an article from Mike Brittain
that was -really- like a sort of new bible to me: clearness, vision, value. The guys there actually built a service from scratch with no (or little, since at least the infrastructure management isn’t hosted by the cloud) up-front investments ending up with a “twenty minutes” service start up time Web hosting infrastructure. WOW amazing.
Think about managing more of one of this services in a complex service infrastructure. Think about managing two, three… ten. At this point we will need clouds to inter-operate. And i don’t think that the only solution is to rely on integration efforts made by cloud providers themselves. Just in case they are not going to comply we will create distributed middlewares, built with federations paradigms, and we’ll deploy them on cloud, possibly :).
At the end, I’m quite sure that a transactional based models will become more and more adopted and things like ACID
, given that me made some efforts to adapt it in a distributed fashion, will be dusted out rapidly.
If we couple this with a new awareness of what are the core requirements. I we succeed to made our:
“…huge step to recognize that it’s time to get over the bias of applying so called “infrastructure 1.0″ requirements to the rules of engagement in the cloud by recognizing that many of these capabilities don’t exist in the enterprise, either…”
– thanks Chris hoff
for this concept I’ve stole on RationalSecurity
– we can achieve sensible results in terms of delivered value and Time To Market.