Modular Architecture And Coopetition
Thinking beyond a single vendor model in services world
Recently, I had a discussion on Twitter with my good friends from Apprenda, a PaaS vendor supporting .NET and Java, about a blog post they published. In the blog post, Sinclair Schuller, CEO of Apprenda, has argued that a vendor who has had success in the app server market cannot effectively offer a PaaS solution with a modular architecture (which requires cooperation with their biggest competitors). He argues that business realities will push the vendor’s PaaS to favor their own app server over competing offerings. He quotes Red Hat’s OpenShift (Disclosure: Red Hat is my employer) as an example where it will still be a JBoss PaaS in spite of the plug and play modular architecture. He was wondering aloud whether one could run Red Hat OpenShift with IBM WebSphere as the application server effectively and whether they will be able to get support for it.
In the past, enterprises preferred a single vendor to provide the entire stack because organizations wanted a single throat to choke for their support needs as well as to avoid licensing complexities. This lead to monopolies, higher costs and lock-in. IT didn’t have any better option but to dance to the tunes of big vendors. This changed completely with the introduction of open source. Open source had collaboration built into the DNA and we saw different projects coming together to work more collaboratively.
Let us examine the simple case of Desktop Linux where core kernel was packaged by third party vendors or community with GNU tools along with other bigger projects like GNOME, KDE, etc.. One could easily dismiss this by saying, Desktop Linux was a niche geek toy and doesn’t meet the rigorous needs of enterprise. Well, I would then like to present the case of how Linux server offerings from various vendors worked in the enterprise space. These offerings, including RHEL from Red Hat, not only supported other open source projects like Apache, MySQL, etc. but even proprietary software. I have come across many instances where enterprises use IBM WebSphere on RHEL. Oh, by the way, none of these organizations complained that it was difficult for them to get IBM to support them on RHEL. If you look at the architecture of OpenShift, you can see that the containers use RHEL (both online and enterprise editions) underneath. Support for app servers like IBM WebSphere in the case of OpenShift is no different from what we had in the traditional IT world. Any argument otherwise is plain FUD.
The trend set by open source has played a significant role in defining how business is done in the services world. We are seeing more and more vendors using a more modular architecture. I am seeing this trend not just on SaaS and PaaS but also on IaaS. The modular architecture in their platform allows these vendors to reach out to partners to offer additional features. Such an approach not only frees up expensive resources for these vendors but also allows their users to use different services (or software) to meet their exact needs. The value proposition offered by the modular architecture is huge because gone are the days when buyers were forced to adjust their needs to meet what the vendor offers. Now, they could easily tailor their platform to exactly meet their needs without any huge investment.
In the services world, a modular approach with a partner ecosystem is the norm rather than exception. Yes, we are seeing some vendors going behind the noise they could create with press release whenever they onboard a new partner. Yes, I agree that the value offered by the partners matter more than the number of partners a vendor could have. However, it doesn’t take away the fact that the real value for the buyers lies in the modular architecture with a good partner ecosystem to meet their needs. Open source has clearly enlightened the modern enterprise buyers on the business value they get from a platform that matches their needs exactly. The modular approach in the services world is giving them exactly they want. The services world has another advantage over traditional IT world. The integration of third party services is much more seamless than ever before and are well optimized in terms of performance.
The enterprise market has moved from a world of competition to a world of coopetition. Gone are the days of single vendor having monopoly in the market and the trend these days is about various vendors cooperating with each other while also competing in other areas. If Microsoft and Oracle, two of the biggest vendors from the traditional IT and whose philosophy is to offer the entire stack in order to own their customers, can come together to co-opete, I don’t see why other vendors cannot do that. Even though coopetition is the norm among startups and smaller vendors, it is slowly becoming the smart way of doing business for even the biggest vendors.
Before I close out, I want to highlight that I am not dismissing Apprenda’s ability to support middleware from different vendors. I am just arguing that it is plain FUD to claim that Red Hat or any other vendor who was successful selling app servers cannot compete in PaaS. I am also not dissing out integrated stacks. I am just arguing that offering extensible platforms is fast becoming the norm and enterprises are pretty comfortable buying various additional services from third party vendors to customize the platform for their needs. Coopetition will be the norm in the enterprise market and there is nothing to FUD about it. #justsayin
