I have to say, “disruption” is one of my least favorite words. But stay with me, because what is happening in the cloud business applications industry right now really is that. Because it turns out what some have called the last great computing architecture for business apps isn’t necessarily the last. There is another cloud wave coming, and it promises to disrupt todays cloud vendors to the core.
Let’s Start by Going Way Back
The cloud in its current form has essentially experienced two recent technology waves. The first in the 1990s and early 2000s which was essentially hosting.
While hosting essentially dates back to the 1960s, led by Big Blue and other providers, in the 1990s vendors were more efficiently able to deliver their applications as a service. The change was enabled by the more ready availability of third party hosting infrastructure and cheaper computing.
In those days, the vendor might host using their own internal data center – with a significant capital cost and risk, or better, use one of the early Application Service Providers to host the application for them. In this way, the vendor was able to offer their application, without the overhead of building their own data center. As the costs of using third party data center providers dropped, more on-premise application vendors were able to offer their existing on-premise apps through this approach - extending all the way to today.
Yet, because many of these applications weren’t really designed to be hosted, they were often limited compared to their on-premise counterparts, and were onerous for the vendor to manage – lack of virtualization technology, or multi-tenancy often meant the proliferation of hardware and application instances required to support customers, while the user experience was often subpar as the application wasn’t typically 100% designed as a web application.
For the sake of simplicity, we’ll call these solutions “First Generation Cloud”. And as I‘ve pointed out in previous posts, it’s very hard for a first generation provider (and there are many around today) to transform their stack while carrying over their existing customers and success to the next wave...
The Birth of the Real Cloud – the Second Generation
The late 90’s saw the birth of the real cloud in the form we see it today for B2B applications. That is, business applications designed from the ground up to run as a service, over the Internet. While essentially still using the same physical delivery mechanism as the first generation – using third party data center facilities, the real cloud saw vendors building applications that were simply much more efficient to deliver at scale. They employed multi-tenancy -- enabling a single application instance or codebase to run many customers. This innovation made efficiently onboarding and maintaining thousands of customers operationally feasible. Further, because they were designed for the web it meant a user experience that was 100% web based – so accessible from anywhere.
These vendors built their new applications on the technology/delivery stacks of choice for the delivery of software as a service at the time: They ran in third party data center facilities. They built and maintained the hardware and application infrastructure themselves – buying the hardware racks, configuring clusters, capacity planning. They used the choice operating systems, and most trusted scalable (and expensive) database and middleware technology available, like Oracle, to ensure they could scale up and out. Typically these companies invested and maintained significant Operations organizations - to monitor capacity, security, and unscheduled downtime, manage and create new clusters.
The barrier to entry for any real-cloud vendor using this generation of stack, was still exceptionally high – they had to work with (and pay) a data center provider, buy hardware, procure an application infrastructure stack, and have talented engineers build multi-tenant codebases right down at the database level, while hiring an Ops team to manage security and availability. Expensive.
And because these applications led from the front in the cloud revolution, they also built a lot of the “context” application technology themselves from the ground up to support their "core" application, because other web services simply didn’t exist at the time. Basically, you pretty much had to build everything for your application, soup to nuts. There were little off the shelf cloud application components and tools that you could use could augment your app to easily add value.
Enter the Third Wave
– The Commodity Applications Cloud
While the first decade of the 2000s saw the rise of the cloud, the second decade has seen the explosion of the cloud. But this isn’t just a fundamental change in software buyer behavior. The last few years have seen a dramatic acceleration in both how fast new cloud vendors can deliver apps to market, and how quickly they can deliver powerful value and world class functionality.
The first driver is that the barrier to entry to deliver a cloud app has dropped enormously. The rise of Amazon Web Services, now a $5BN business for Amazon, as well as other Infrastructure-as-a-Service platforms like Microsoft Azure eliminate the need for them to do deal with costly co-located data centers, and the need to provision, buy and depreciate expensive hardware. You can get started at almost no cost and no economic risk.
The third-generation get better elasticity, and avoid the need to buy infrastructure resource ahead of peak periods. And they get core operational security and availability out of the gate. This is everything that the second-generation cloud provider had to build and manage themselves. It means they can spend valuable resources on apps, not infrastructure.
But there’s more. The cost of application infrastructure is dramatically less – with databases like MongoDB, PostgreSQL, OrientDB, ArangoDB, or MySQL that are feely available, and provide the right level of functionality, at a dramatically reduced cost than the traditional database providers used by the second generation. The same goes for the rest of the application infrastructure stack too. Less cost, and also simply more accessible for developers to get started with. And modern application frameworks like Ruby on Rails make it easier than ever to develop and iterate web applications – which their second-generation counterparts often aren’t built on.
Finally, the third-generation have the benefit of being built in a world where everything is now as a service, and service enabled. Need to add analytics to your application, why build it? How about Amazon QuickSight or PowerBI? Need to add predictive analytics, why not Microsoft Azure Machine Learning? Need to add an Excel interface? Why not simply plug into service enabled Microsoft PowerQuery? Need to roll up data? Maybe RedShift? Want workflow? Why not incorporate KiSSFLOW or any number of other workflow tools? Need data integration, why not simply plug into SnapLogic or MuleSoft? The list goes on and on.
What’s changed is what the second generation previously had to build (and now maintains) themselves is now available off the shelf for the third generation. Simply, the new breed of commodity applications cloud provider can innovate faster than their second-generation predecessors. There's definitely some irony here, for sure.
What It Means
This transformation is structural. In my opinion, it’s just as hard for a second-generation cloud provider to move to third, as it was for the first generation cloud to make the shift. The competitive implications are substantial.
It means an increased velocity of new providers entering existing software application categories with new cloud applications, built using commodity infrastructure and incorporating rich application services provided by other vendors. It's simply easier to enter a market than it it ever was before, and enter with a strong, competitively priced offering
It means the third generation can attack their bigger second generation cloud competitors with applications capability built on the shoulders of others, while focusing a much larger percentage of their R&D on their vital areas of differentiation, rather than building and maintaining technology and application capabilities that have essentially been commoditized.
And ironically, it offers first-generation cloud providers a way to get back in the game - it's never been easier to get into the cloud.
It means more competition. It means more choice for buyers. It means more innovation. And yes, it means more disruption.