“People who are really serious about software should make their own hardware.”
Something really intrigues me… In the not too distant future, assuming a majority of organisations are hosted on service provider infrastructure (ie XaaS) rather than running in-house infrastructure, does this imply the number of customers for hardware (compute/storage/networks) will significantly reduce? Service providers would obviously be major buyers, but who else do you envisage would be buyers of hardware?
The reason I ask is because I’m intrigued about the impact this question has on OSS as we currently know it. Does the customer base for traditional OSS (ie inventory, alarms, performance, etc) also shrink down to the service providers that buy the hardware? But in turn, this leaves an opening for a different type of OSS (ie managing apps, content, cloud infra management) to evolve to needs of the enterprise customer base.
The simplified diagram below demonstrates the concept with customer-facing (above the line of complexity abstraction) tools and SP-facing (below the line) tools:
The Service-Provider-facing (Below-the-line or BTL) OSS will increase in complexity because there are more layers to manage in a virtualised environment. I won’t get into that in today’s post, but will undoubtedly be discussing it in upcoming posts. You can also see more about this from an earlier post about the components of SDN, NFV, MANO and OSS.
The above the line (ATL) OSS (if OSS is even the right name for it) will need to provide a drag and drop interface that allows an organisation’s designers to easily add servers, storage, networks & connectivity, security devices, applications, etc. This interface will also allow the customer to see network diagrams, utilisation, service reliability against SLAs, events, order tracking status, etc.
Now you might be asking why the ATL OSS is not supplied as a portal by the service providers. Actually, that is one possibility. However, the major benefit of providing this as a third-party tool is to allow a customer to independently pick-and-choose-and-migrate capabilities/offerings from more than one service provider.
The customer-base for this type of tool is any organisation that relies on a communication network and/or hosted services (ie massively larger than the current pool of potential OSS customers).
At some point in the future, I’d expect that the customer-facing interface could be simplified to the point that almost anyone could build a technology environment to suit their organisation’s needs across multiple service providers, not just the IT/network gurus of today.
There are only three ways that this level of simplification can occur:
- The options within the underlying infrastructure / services have to be heavily simplified compared with the massive amount of configurability options offered today AND/OR
- The abstraction of complexity (as shown in the red box in the diagram) becomes an interfacing master-work AND/OR
- Managing the infrastructure cloud would need to be done using entirely different approaches than are used by current OSS (see more on this complexity here)
EDIT. The Cloud Portfolio Management platform offered by Rightscale has been brought to my attention by a colleague as an organisation that’s building out capability in this space.