As a consumer of the cloud, riding on top of the commodity hardware is a layer of software defined infrastructure. The logical manifestation of infrastructure removes much of the need for technical architectures. Standard patterns or reference architectures are leveraged to bring the physical infrastructure to fruition. The fungible nature of software defined infrastructure. Much of the architecture to support applications becomes a matter of scripting so it can be recreated on the fly. Integration becomes a logical exercise and forgoes much of the traditional technical architecture. A paradigm that subverts the architecture is microservices. These are, for the most part an integration pattern.
Contrary to the technical layer of the technical stack, the areas of business and data become more important. I believe this is due to the shift away from datacenter centric architectures, the fungibility of software defined infrastructure and dev-ops deployment patterns. Things like cost management become more important since commoditized resources are rationalized through an abstraction layer. Patterns like turning off instances or even deleting them until they are needed in the future is a new mechanism of cost savings. The business layer also becomes more interesting as we expand into the cloud driving the need for more clear ownership of our cloud infrastructures. No longer can we rely on location to ascertain ownership especially for large organizations where hundreds or even thousands of applications exist in various forms. Without that clear ownership, we create gaps in how we run, support, develop and maintain those applications thus creating a potential risk to the customer and business alike.
The data layer is another area that is critical to operations and should have more scrutiny in the cloud. In my experience the cloud offers easy ways to store and manage data, however there are considerations with regard to location and sensitivity. Inherently your data is exposed to the world, however security architectures implemented in the cloud serve to ultimately protect your data. This is why gartner alludes to a security architecture. Aside from encrypting and securing your cloud data, there is another concept called data sovereignty (see http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=6848696). This is the idea of ownership of that data and where is it geographically located. There is governmental oversight over the data and therefore patterns must be upheld in how that data is handled, catalogued and tracked.
In conclusion, the technology stack is changing and I would estimate there will be profound impacts on how EA comes together to provide value to the organization. EA is traditionally a “dark art” that has not quantified its value through decision and in practice is becoming more of a governance function. For the practice to stay relevant, we must shift our focus around technology and reinvent EA. Perhaps augmenting the practice mirror the principles of Bi-Modal IT would help to outline the current and future landscapes.
