It’s been a few months that I’m thinking about “cloud” business. Even if the concept of SaaS is around since early 21th century thanks to visionary people like Tim O’reilly and others, if you give a look to this, it appears very clearly how, instead, “cloud computing” concept seems something that no ones comprehends completely or at least a container of lot’s of interesting things like *aaS, paravirtualization, grid computing, thin clients, etc…
Anyway, let’s try to go beyond the complexity of terms and to focus on a couple of ideas: SaaS on one hand and PaaS/IaaS on the other hand. With SaaS I actually mean a provider creating a software and then suppliyng this software “aaS” via the interent to the customer base (a good example could be SalesForce.com, a truly pioneer company or WordPress.co – even if in this case the end user is more a single person than a company). On the other hand we have PaaS/IaaS offering providing the customers with, generally, lower layers empowered by runnig on top of a cloud (data center :D), easy scaling and exposing (sometimes exclusive) API’s or performance enhancements that can differentiate from what we formarly know as an hosting service.
For what regars the SaaS/IaaS area I completely agree with Simon Wardley at Gardeviance when he says:
“The real battleground for the “cloud” has always been in building an ecosystem around the framework layer of the computing stack.”
So is a matter of enrich a traditional paravirtualization approach with new and powerful tools (like open API’s could be). Isn’t it?
Given that short distinction (I know maybe you could have different ideas) lets try to focus on the opportunities that the gives to different players on the Information Technology market.
From the End user company perspective, I believe that one of the main drivers of IaaS,PaaS is truly the cost. Who would’t agree with having the IT resources commoditized and available on-demand with no or small configuration effort? surely I’ld generally do.
Unfortunately, more or less in line with what Dennis Howlett was saying some days ago on ZDNET:
“Putting my buyer’s hat on, I want my computing infrastructure to be available on demand at the lowest possible cost. I want to drive those baseline cost efficiencies into as many of my applications as possible, but not at the expense of sacrificing or endangering security, or my ability to run a compliant set of applications. In many scenarios and especially those that are regulated, operational code ‘ownership’ is important.”
security and compliance constraints can show up really soon an made you reflect more on the choice.
Some days ago, when I was talking about this with Andrea I realized that, in fact, core services/applications are typically difficult or unworthy to “cloudized” (being erogated via a PaaS/IaaS) due to a bunch of reasons: security and privacy of the data, responsibility about system availability, code ownership and disclosure bonds.
In fact is very common that the bigger part of a solution’s (or architecture) TCO is merely represented by the data tier in itself. Normally, the data tier also is the tier most subject to constrictive security and availability/continuity rules (since, indubitably no system works without the data).
These few reasons actually could represent a blocking issue especially for customers dealing with critical infrastructures (eg: defense or public safety sector) that, on the other hand, would tipically require an adaptive infrastructure able to manage peaks and, sometimes, bursts.
Even if there is a very wide set of end customers in the Service sector that whould be interested in approaching PaaS/IaaS (having a set of less restrictive continuity/security rules) for example TELCOs, Post&Logistics or e-commerce, we must consider that, very often, such kind of customers have a pre-existing infrastructure that have been consolidated in time, often thanks to relevant investments in OSS/BSS systems and often including a relevant level of cooperation among different parts (with a strong SoA orientation) and managed by large teams with consistent knowhows and values.
Instead of focusing on the biggest customers out there, IaaS/PaaS players should concentrate on emerging companies, that from the start, should build their own infrastructures simply seeing IaaS/PaaS as one of the choices, one of the available tools. And, to be true, it’s merely what I see offered right now by MS Azure, Amazon Ec2 or Google App Engine. A powerful and appealing offering in this area should be based on openness (API’s, standards) and ecosystem (spread knowledge, create new skills for developers, engage SIs and other player as partners).
The opportunity for promoting new views when buildling an infrastructure are here, now we wait for the community to get interested, evaluate, promote, use and finally make it a standard solution to standard problems.
Obviously the ecosystem is made in itself by very different players. If we put an eye to potentially involved third parties (ISV, product companies,consulting and oustourcing firms, System Integrators) this effectively represents a big opporunity, and is a matter of vision and promptness to catch it.
I’m not so aware about US situation. I’m quite more informed about EU side, since I’m working for an European consultancy firm that is actually a leader in innovation and very well distributed across all Europe thanks ot its subsidiaries.
I could say that the overall openness to investments in Europe is not so high. IMHO this scarce availability to invest have often prevented good ideas, innovative services and propositions, reliyng on a networked infrastructure, to have their chance on the market. In fact, upfront investments, often keep the ideas sleeping until the first customers shows up and, very often, gets parially and implicitly charged of part of the CAPEX.
We can think PaaS/IaaS and Open Source software to be used from third parties to create SaaS or other offerings with lower upfront investments needed.
It will surely represents a key success factor, the ability of PaaS/IaaS providers to ensure little or no downtime in the future and to avoid that things like this show up again in the future. In fact 8 hours downtime will prevent not only critical infrastructures to exploit the IaaS/PaaS paradigm but also player used to think by a money per transaction per second paradigm.
At the end, I believe that “Cloud computing” actually represents another innovation driver, as Open Source is being more and more.
Is there an opportunity out there. We need only to catch it. And, since we’re talking about clouds, we should start thinking that sky’s the limit ;)