Sometimes I love reading commentary from El Reg on IT and technology. Delivered with just the right amount of bite, their bark is often right on target.
Other times, though, they miss the mark in their biting commentary. In a diatribe on cloud early this summer, “Cloud is a six-horse race, and three of those have been lapped”, they lament the results of the latest Gartner MQ. The latest analysis dropped multiple cloud providers based on market requirements. This led to a rather scathing response from El Reg, despite the fact that in this case, at least, the “Gartner mages” are absolutely right.
“Customers now have high expectations from their cloud IaaS providers. They demand market-leading technical capabilities — depth and breadth of features, along with high availability, performance and security,” wrote Gartner’s mages. “They expect not only ‘hardware’ infrastructure features, but also management features, developer services and cloud software infrastructure services, including fully integrated PaaS capabilities.”
The latest MQ dropped eight different IaaS providers, leaving AWS and Azure to reign with a trailing Google as leaders, and marking Alibaba Cloud, Oracle, and IBM as niche players. No challengers. No visionaries. Just leaders and well, not leaders.
The thing that got my goat was not Gartner’s analysis. It was El Reg’s dismissal of their criteria and its heavy reliance on what goes by many names, but can be summed up in part as “application services” along with developer services and integration capabilities.
Oh noes! Who would have thought that enterprises are demanding – and choosing – IaaS providers that offer enterprise-like features and capabilities? It’s inconceivable!
Let us return briefly to 2013 and a discussion spurred by an article detailing what cloud providers were “getting wrong.” One of the things they were getting wrong (and some still are) is the failure to differentiate through application services and integration.
From my 2013 screed:
IaaS differentiation is not going to come through more varied instance sizes and configurations or through price wars. Enterprise customers understand the value - the business and operational value - of services and pricing is less a deciding factor than it is just one more factor in the overall equation. The value offered by pre-integrated services that make building a hybrid cloud easier, faster and with a greater level of reliability in the stability of the service - such as would be offered by such services - has greater weight than "is it cheap."
When you move an enterprise application to the cloud, you aren’t moving just an app. You are moving an architecture. An architecture comprised of many (an average of sixteen) different services. Some of those services can (and should) be replaced by application services native to the target cloud. There’s the obvious ones – any service that would have been “shared” in the corporate network is best served by the cloud provider. That’s primarily your network and application infrastructure, but also includes logging.
Others application services can (and should) be migrated in order to maintain compliance and consistency in policies – especially those policies that are handcrafted on a per-app basis to match corporate security requirements. Web application firewall policies are notoriously app-specific and are the most likely to migrate from on-premises to the cloud. The investment in policy management alone is worth the migratory effort.
Regardless, this practice presumes that similar application services (and integration) exist in the target cloud. Or that the application services used on-premises can make the transition to the cloud without introducing excessive overhead.
This mix-and-match, a la carte approach to application services is one of the reasons we’re seeing the rise of per-app application services. Because cloud needs enterprises to sustain their growth, and cost isn’t the driver it used to be. Differentiation through integration and application services is the way cloud providers are going to attract and retain enterprises.
But they aren’t going to offer them all and the average number employed by organizations continues to rise. That means orgs need to break up the monolithic delivery architectures of traditional data centers and adopt a per-app model that enables a plug-n-play approach to choosing the application services needed to maintain availability, improve performance, and sustain a secure environment.
I’m sorry to say El Reg myopically missed the mark on this one. The analysis is spot on and only off-handedly recognizes the enterprise demand for differentiation. It’s been a sound bet that at some point, the costs couldn’t go any lower and cloud providers would be forced to find new ways to attract customers. Application services has always been the best bet for providers to differentiate.