One of the perhaps somewhat counterintuitive challenges with at least the initial stages of adopting Fusion Middleware is the fact that there is too little work in terms of administration.
On the one hand, Fusion Middleware administration entails quite a bit, starting with WebLogic Server
and typically extending to one or more FMW components:
all of which the administrator – or rather the administration team – needs to deal to with. Typically even around the clock to ensure the availability required by the business.
On the other hand, the actual workload for FMW administration for a small number of applications, services and processes does not justify a dedicated resource. This proves a serious problem for many organizations: 24/7 availability requires 3 FTE while the effort is on average less 0.5 FTE.
Organizations can adopt several strategies to address this challenge, as is illustrated in the next picture.
Middleware administration can be combined with responsibilities for other components in the stack. This can either mean moving up in the stack: more business oriented administration of the applications and other components deployed on the middleware or, more common, moving down to also take care of the database.
An other approach to ensure a full time workload for the middleware administrator is to engage the administrators with the development teams and pre-production environments. This has many advantages, including better and more production-like environments for the development teams, accelerated built up of experience for the administrator, much smoother hand-over from development to production and early focus on non-functional requirements and deployment.
Another obvious way – although in reality hard to realize in some formal or political organizations – to broaden the responsibilities of the middleware administrators is to have them look after environments and applications in other departments or organization units than their own. Extending from that idea it might even be an option to share administration staff between different organizations. Of course these last two approaches are very similar to the idea of the cloud – but instead of sharing infrastructure it is sharing human resources.
Additionally, most of the day to day operations is fairly run of the mill work that does not require very deep knowledge or long experience. However, every now and again, a special task needs to be performed or an urgent issue resolved that does require such knowledge and experience. It seems most realistic and efficient for organizations – at least until they reach a certain critical mass in terms of [Fusion] middleware workload, knowledge and experience – to have external specialist on call to help out with these out-of-the-ordinary tasks. Such specialist support can be captured in part-time presence and SLAs (Service Level Agreements) along possibly with some remote monitoring. The external FMW specialist can also coach the internal staff, helping them gain knowledge and confidence.
