Server Management – Part 2 (sort of)
Posted on | May 19, 2009 | No Comments
So the discussion on the LinkedIn forum continues on this topic and I was compelled to write a follow-up to the post I included in my brief article I posted here. Naturally I felt inclined to add the meat of that discussion here by way of another article.
—
It is hard to provide a “package” that addresses a customer’s needs without first engaging with the client to properly ascertain what they want to do and what are their limitations and expectations.
As I said, if you allow the customer to select from a range of packages, unless they know preciously what they need and what each service entails etc. they will most likely expect more than you offer or get more than they need. In either case neither is good business. In my opinion managing someone’s server is not as easy to define (or sell) as a t-shirt where all that matters is colour & size.
This is were, I believe, virtualisation can make the client experience and the provider service a better match. This is especially true if a client isn’t sure of their needs and more so to the benefit of the provider if they use the right virtualisation solution. While it can simply be a virtual environment (VE) on a single server there is no reason it can’t a VE on a grid or cluster, thereby eliminating many single points of failure. This would give the provider an easier platform in which to control access, drag and drop OSes and near instantly provide resources as required. It is also a standard platform which is the same for each of the provider’s clients and therefore much easier to support.
Thankfully many virtualisation systems are already on the market that can cost effectively do this and are mature enough to be seriously considered for most customers.
Ultimately it comes back to what the clients expects to achieve from the server, what business services it will preform or provide and what fault tolerances are appropriate… in this case one size certainly does not fit all and there are many additional service to consider. Such as back-ups.
I question whether this can be done effectively if the server being managed is not local, to the managed server provider, unless that server is virtual and is not on a single physical server. It is far easier to resolve problems when you can simply mount the server in another grid or stop it add resources and then restart it. Patching becomes easier if the managed service provider maintains a test environment (again virtual) where updates can be evaluated before going onto production systems.
While this kind of redundancy maybe overkill for most requirements it is vital to an enterprise operation and if it can be provided at a price point that is both reasonable for the client and profitable for the provider then there is every reason to consider it. Doing so is the very definition of ‘proactive’ and I would add it is what the client would do as only they know what their business truly means to them. Business continuity is always the paramount concern of any business. Such safe guarding becomes more important when the client chooses to outsource their support.
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_e.png?x-id=d961bb12-ea87-420f-a4ed-e374fc8b9fd4)
Tags: Business > Customer > Facebook > help wanted > LinkedIn > Managed service provider > Meat > MySpace > Networking > Outsourcing > Single Point of Failure > Social media > Television > Twitter > Virtual reality
Comments
Leave a Reply