Bernard Golden| Cio
I observed an interesting debate on Twitter a couple weeks ago between an advocate of “enterprise” computing and an Amazon Web Services champion. After it went back and forth I bit, I offered my contribution: Somebodyis using a ton of AWS, and it’s growing like crazy. Listening to this debate reminds me of the Men Are From Mars, Women Are From Venus discussion about how two people can discuss something and still fail to understand the other person’s basic perspective. In the case of this Twitter debate, the discussion failed to address a key question: What are the requirements of the applications running in those environments?
The crucial fact is that those who defend enterprise computing fail to grasp the fact that legacy IT infrastructure and operations don’t address the requirements of new application types that I label the “three M’s”—mobile, media and marketing. These apps are flocking to public cloud computing because they’re not well served by traditional infrastructure and are much more aligned with what cloud computing brings to the table.
It’s critical to understand the characteristics of these applications to understand why demand for cloud computing is in its early growth phase—and why we’re about to see its already rapid adoption accelerate even further.
Cloud Computing and Mobile: A Different User Population
Legacy enterprise applications could be tuned to a couple operating systems and a few browsers. They also had very predictable user populations and use patterns. The emphasis for these kinds of applications vis-à-vis infrastructure is to implement a static environment and make it difficult to modify.
Mobile applications, on the other hand, are a very different. They run on a bunch of different devices, which increases the combination of interfaces that applications need to be able to support. Moreover, companies often provide API interfaces to their applications to enable independent developers to create applications outside the purview of the company’s own IT organization—the company won’t even know what devices are going to be in the user population.
The growth of APIs is one of the real underreported stories of the past couple years, but it’s huge and very much driven by a mobile application world. (The upshot of this is that mobile applications pose significant challenges to the design and operation of legacy applications.)
A further issue is that mobile applications are much less predictable in terms of user populations and application load. All it takes is a prominent Twitter user or Facebook celebrity to highlight the application and make traffic skyrocket 10 or 100 times. When high load variability meets legacy infrastructure, something’s going to give—and it’s most likely going to be the infrastructure.
