Modernise and prosper: It's time to imbibe server orchestration

Live long by transporting what you know to a virtual environment

Digital fuel for you, too

This request can then be submitted and a fully built virtual server should hopefully be built and deployed as requested. Other clever orchestration workflow features include the ability to set limits so that the requester can’t get too greedy and request multiple terabytes of disk space for a basic server. If a server request (meaning things like CPU, RAM or disk space) is above a pre-determined limit, logic can ensure that the request can be sent to a more senior person for approval. Such logic can be very useful if the company in question has a chargeback model at work.

This is partly why vendors such as VMware are gobbling up companies such as DigitalFuel, to incorporate accurate costings for the virtual machine costs based on real-world hardware costs and overheads. This is especially useful in a chargeback model.

Whilst this may sound like something that a basic portal could do, the real secret sauce is that in a well-created orchestration system, the back end of the orchestration system will also handle jobs such as adding the server DNS entries, IP address and CMDB (system configuration databases). The orchestration logic is simple enough.

The interesting part is how orchestration logic is married to the business logic. The two most difficult parts of setting up an orchestration system are getting the system to work well with third-party systems and getting the business logic just right. Once done, however, it pays dividends in cost of deployment and lower management overhead.

Hopefully, by now you are convinced that orchestration is a useful tool. The big question is how this affects you, as the system administrator. In small companies that stand a handful of servers per month, not a lot. Those of you that work for service providers or larger companies, however, need to get on board with orchestration.

Working within an organisation, I have seen jobs that consumed most of the day for a staff member and two at peak times (deployments) become fully automated. The workload dropped by a good 80 per cent. There will, of course, always be exceptions to the rule, which are not automated for whatever reason – usually involving very complex and specific designs, as well as deployment of pre-packaged machines, such as OVF deployments.

So, my fellow admins, it is time to get ahead of the curve rather than be left behind. There is exceptionally good money to be had if you are proficient at it. What makes getting good at orchestration a really good idea is that it is no longer a purely administrative task.

A proficient orchestration administrator needs to understand in detail not only the hypervisor but also the automation and integration with third-party applications. On top of that, the ability to grasp business processes and convert them into logic routines is critical. Also, there is the need to be proficient in scripting and API use to make full use of it. To be fair, you can get third parties to do this – but that’s cheating.

How big is this? Put it this way: a certain tier-one vendor which possesses a name-your-price mentality has already tried to hire one of our internal cloud and orchestration gurus on several occasions. That should give you an indication of the level of demand for those with skills in orchestration that are married to the cloud.

Administrators, this is the way of the future. No longer will it be a disjointed piecemeal approach, but instead it will be a self-service environment with a catalogue of services which can be selected and purchased.

So what orchestration tools should you look at? The short answer is that it really depends on your environment. All the major vendors will have their own orchestration suites and methodologies.

What happens next is up to you. ®