Why PaaS is the New Black

I had very interesting conversations and strategy sessions with my clients lately, and noticed peculiar things in the market as well — all of them around the same issue: Platforms.

No, I am not talking about the shoes of the 1970s we loved so much, I am talking about the intermediate layer of the cloud model – Platforms used to deploy applications.

I wrote about Chatter back in November, and Genesys couple of weeks ago — and in both I expressed my firm belief that Platforms are going to be the issue that defines the cloud in the next few years.  I also wrote about the cloud as being more than simply SaaS (Software as a Service) applications in my 2010 “predictions”.  I truly believe this is going to be the issue that will define the next generation of enterprise applications (this whole SCRM v CRM v Enterprise 2.0 v Social Business has limited future — and even If I am wrong, they still need a platform to run on, right?).

Here is where I stand on this, an unedited version of my brain right now (careful, may scare you).

There are three layers to the cloud: IaaS (infrastructure), PaaS (platform), and SaaS (applications).  They all ride on a universally known and publicly available network — they need it to exist (note: thus the impossibility of the “private cloud” — sorry, pet peeve).  This network is controlled by — well, let’s face it Cisco and a couple of others that have some products here and there.  OK, mostly Cisco.  This “Mostly Cisco Network” supports the infrastructure layer which essentially handles the communications between the network and the platform, while providing some services (authentication, security, encryption, integration and links to database and legacy systems, and common protocols among others).

Right on top of the infrastructure is where the platforms live also interacting with the layer above: the applications.  The platform is the management layer that connects the infrastructure with the logic and presentation layers provided by the applications.  Here is where something like a community, a knowledge-base, and a rules engine (as examples) would exist in a cloud environment.  Platforms provide an answer to the application on whatever information they needed, with the infrastructure and network supporting them.

Finally, the applications – the stuff that truly, honestly is the easiest of the three (complexity decreases as you climb the three layers of the cloud – or the seven layers of the OSI model in which it is modeled).  As I used to say about survey software, anyone with a garage and a couple of weekends can build a cloud applications (as long as they have the platform and the infrastructure in place — otherwise is not a “cloud application”).  Maybe more than a couple of weekends, but you get the idea..

This brings me back to my original point that platform solutions seem are emerging.  A platform is what would make a call center (OK, contact center and we can accommodate several channels) flexible, dynamic, and able to add a new channel (say, like Social Channels) with relative ease.  The social networks we talk about so much are all platforms (yes, Twitter, Facebook, Communities, etc.).

Why am I bringing this up?

I am sensing a rising problem: interconnecting the platforms.  While ideally and in theory this would be handled by the cloud as long as the platforms support an underlying infrastructure, this is not the driving force for the design of new platforms.  I am seeing platforms that MAY be open and easy to leverage and integrate, but with zero effort spent in trying to figure out how these platforms can and should work with each other.  I am seeing half-baked efforts at platforms that don’t consider integration with infrastructure and other platforms as vital. This is not only bad, it makes the platforms not cloud-compliant and thus not very useful in the long-run.

I am trying to make sure that the new platforms are indeed open, integration-ready, and cloud-compliant.  I want to raise the flag early on so we can actually leverage them, and make application development easier while making the problems they solve more complex.  I want to make sure that this time around the cloud actually has staying power since my poor heart cannot take another CORBA-style disappointment.

Got it?

What do you think?  Am I asking too much? Is the cloud even possible? Are platforms going to be the big thing for 2010?  Would love to hear your thoughts on this…

16 Replies to “Why PaaS is the New Black”

    1. Ryan,

      Yes, absolutely, 100% you cannot just say web services. Actually, you can — as long as you have the proper infrastructure (the three layers of the cloud) leveraging them in an open, dynamic, and flexible model to develop against.

      Thanks for the kudos, but more important for validating my point that this is something that others are seeing / experiencing.

      Thanks for the read, and welcome…


  1. Dear Esteban,

    Rightly sad. There are interesting benefits in deploying your platforms and applications from the cloud, be it a public cloud, a private one or a combination of both. Either way, for any corporation using IT there will always be a need for integration between the various applications and data sources.

    By experience I know that Salesforce is very keen on making integration very easy. I hope other are vendors have been doing their homework in this area, to. Up until now I am amazed about the still very small number of IT consulting companies providing integration services in this area.

    Perhaps you just discovered a huge opportunity for IT service providers.

    Best regards,


    1. Robbert,

      First, when you say “I hope other are vendors have been doing their homework in this area” in regards to integration, all I can say — that is my most sincere hope. Without the easy, open integration provided by leveraging the three layers of the cloud, there is not much benefit of using a cloud (other than marketing prowess of not having to call it on-demand). I think that this is the critical aspect to understand — and why I wanted to highlight it ever so.

      Second, if this comes to be a huge opportunity — I am going to go back and claim some royalties on that work 🙂 I am hoping we don’t end up with that model, since the applications and platforms will be developed to leverage the other layers in the best way possible and the role of the SIs won’t be building point-to-point integration adapters like they do now.

      Thanks for the read.


  2. Not asking too much – v legitimate concern.

    I’ll up the ante though:

    This time around, lets please not have the platforms/infrastructure figure out how people need to work together. One ERP-gestapo’d interaction model per century is enough. -)

    Lets make sure that these platforms/infrastructures are open (embrace APIs and Plugins) and let the coming convergence model around customer/relationship centricity dictate where and how the glue needs to be applied.
    .-= Sameer Patel´s last blog ..The iPad: The Read Web is Ushered Back In =-.


    1. Sameer,

      Now, if you are truly going to push for a cloud-centric model the least you could do is warn the vendors what they are going to have to deliver. Oh, wait… I am going to do my part of pushing for this “glue” to bring it all together to be open and to follow the rules of the cloud model as much as possible. I hope others do the same.

      And, btw, I agree with you that we have had enough of this disjointed, disparate models — I sure hope that we can learn from our errors and focus on developing to the standards we set this time around, and it should be easier to get to where we are going.

      Also, killer line you got there “…let the coming convergence model around customer/relationship centricity dictate where and how the glue needs to be applied”. Nice.

      Thanks for the read.


  3. Esteban

    Interesting post. In a follow up, any chance that you can provide a diagram that lays all of these out? Just curious.

    Thanks. I’m following you on Twitter now.


    1. Phil,

      Been working on that “simple chart” for about 4 months now 🙂 It is part of a report called “The cloud for CRM Dummies” (working title, be glad to change it).

      Problem is time available… will see if i can take one of those charts and make a post out of it quickly.

      Stay tuned! (thanks for the twitter follow and the read here)


  4. Well thought post, Esteban.

    In all the scary thought of security on the cloud, the open source app call is well thought. I really wonder how the different industries that this platform will have to serve really react. Businesses need to understand the pains of IT but will IT be able to live up to business expectations vis-avis its painpoints? That remains to be seen.

    Though SFDC – as rightly mentioned in one of the comments – is upto it, vendors moving in the direction you suggest will benefit.
    .-= venkataramanr´s last blog ..Infosys at the SAPinsider event =-.


    1. Venky,

      Thanks for the read and comment. I am not sure how the industries will react — but i’d prefer not to ask them 🙂 I’d prefer to do what needs to be done and if someone needs to use stronger security or better collaboration tools they can find it. The main message I had was about leveraging the platforms instead of re-inventing the wheels each time. I sure hope ti happens, as it will make a considerable difference.



  5. Hi Esteban,

    Great post, but I never expect less from you. I have mentioned to people multiple times that without forethought, what is happening in the cloud looks EXACTLY like what happened in the early 2000’s: customers bought PeopleSoft for HR, Siebel for CRM, SAP for ERP, and none of it worked together, so then they used duct tapes and band-aids to try to turn the point-to-point mess into a more manageable hub-and-spoke pattern. The cloud in some ways EXACERBATES this problem, because the number one constraint in the cloud environment is network bandwidth, so data integration, let alone process integration, is even more difficult to execute due to limitations in the size of the pipes between clouds.

    Best Regards,

    .-= Nenshad Bardoliwalla´s last blog ..The Unified Performance, Risk, and Compliance Model – Part IV – Model and Optimize =-.


    1. Nenshad,

      Thanks for a great comment. Here is my position on the issue of database (I have been trying to raise this issue for a very long time now, few understand the issue of data storage and data management, and how complex it gets really quickly): I do believe we are at least 2-3 years from seeing the incipient ideas I highlighted above make it into the market somehow. I have began to see some changes in data storage and data management (including your new venture — congrats, btw) that are likely to deliver value in the near future. If that is the case, while I agree with you from today’s perspective of what data do — but i am going to take a different perspective in 3-4 years.

      I think that Venky talking about security, and you talking about data management in the cloud — are the issues we need to raise now. Alas, I want ot make sure we do more than talk and we get the proper elements in place for everyone to leverage. Possible? Hope so.

      Thanks for an excellent comment, as usual.


  6. Great article Esteban and what a great discussion!

    You’re correct that the platform layer of the cloud will see much development in the near future. I fear that competitive pressures will create a world we’ve seen before… Several major platforms that don’t play nicely together – i.e. Windows vs. Mac

    Microsoft, Salesforce, Google and Oracle are all creating their own platforms are compete head to head in core areas.

    The idea of open standards to allow cross platform connections is nice but, the players in this game have a history of wanting to control their platform to gain competitive advantage.

    Google is the exception obviously but, their market is the small business, not the enterprise.

    Its likely that we’ll have a few years of platform wars and then, when the next big thing comes along and the players need new users so badly that they’re willing to cannibalize their cloud revenue by connecting platforms with competitors, we’ll all get integrated.

    I suppose that in this particular game there is the opportunity that application layer (software) vendors can drive the issue. All platform competitors will want to play with the killer apps.
    .-= Craig Klein´s last blog ..3 Ways to Measure Advertising and Marketing Return Easily =-.


    1. Craig,

      Firsts, apologies for taking so long to reply — just saw this today.

      While I agree with you, in principle, in the standard behavior for the vendors mentioned, I may be a bit more hopeful this time around. If what you say were to carry some weight, then Microsoft Azure would have already become the de-facto platform for cloud projects — which has not. I still believe we can, this time, begin to see some inter-platform operations using standards and methods we can all agree on.

      I actually don’t see Google as the exception, but rather SFDC as the exception to this. Force.com is the most advanced platform I have yet seen and the one that truly adheres to “open standards” as much as we have them right now. If the continued development follows along the current lines, I am hopeful we can see some progress in the next 2-3 years.

      I do totally agree on the years of wars, but I am hoping that a simple, open way to interconnect shows up soon.

      thanks for the read!


Comments are closed.

%d bloggers like this: