Archive for January, 2009

Effective Operational Processes in Storage Environments

You’ve decided which storage technology to use and constructed a suitably compelling business case to convince the Financial Director to part with the cash to pay for it. The supplier resources have been lined up to install and configure the new kit and software and you’ve appointed a project manager and project board members to ensure that the implementation is successful. The project is up and running and everyone in the organisation is eagerly awaiting the benefits that the new centralised storage solution will bring. But hold on a minute – who’s going to look after all this new stuff and what processes do you need to set up to ensure that it all runs smoothly once implementation is complete? As the saying goes, the hard work starts once the project’s finished. Read on to find out how to keep your storage solution ticking over nicely once it’s in production…..

Over the fence it goes

It’s a sad fact that many organisations which have mature, well-adhered to project governance processes and standards and are, as a result, very efficient at delivering new IT projects on spec, on time and on budget, often lose the plot somewhat when it comes to handing these new solutions over into production. With some justification, Operations and Production Support departments often feel like the poor relations within the I.T. division; new systems are often “thrown over the fence” from Development with little thought as to what is required to ensure that they can be adequately supported and operated throughout their working life. Poor old Ops & Support are left scrambling to get up to speed with the new systems and, of course, everyone blames them when things go wrong! At the extreme end of the spectrum, it may be impossible to use the new system altogether as there is no capability to support and operate it – effectively meaning that no value is derived from all the money spent on it. This, clearly, makes little sense.

Think about Production Sooner rather than Later

The first thing to get right, therefore, if you want to ensure that your organisation really gets value for money from the money it’s spending on storage, is to involve Operations & Support from the outset of the project. Think of it another way – until the solution goes live, it’s just spend, spend, spend – it’s only once the thing is in production that it starts to pay back and deliver all those benefits claimed in the business case. Operational and support considerations should, therefore, be addressed from the outset of the project and costs for internal and external support resources, training and maintenance contracts should all be included in the financial calculations for the project.

Look at Your Organisational Capability

If your organisation has never used a technology before, it’s highly unlikely that it will magically have the ability to effectively support and operate it once an implementation project is complete. Although a number of individuals in the I.T. department may have, or claim to have, skills or experience in SANs, NAS, capacity management and so on, and most systems administrators will have set up an HBA card at some stage in their career, such skills are generally more effective if they are available within the context of a defined service offering, process structure and organisation. Let’s look into these three areas in a little more detail.

Define your Storage Service Offering

What does “storage” mean? Well that’s simple….or is it? And does it mean the same to everyone? Does it include local disks as well as centralised storage devices? Backup & restore as well as online data management? All platforms and operating systems or just distributed? For the new service to work effectively, everyone has to be clear what it is and what it’s made up of. A good way to get to this point is to define at a high level what “storage” is and then break this down into component activities to arrive at a service and activity “catalogue”. For example: -

Service Description

“To provide a suitably sized, performant, secure and highly available enterprise data storage platform and ensure that all critical data is recoverable within agreed service levels in the event of accidental or catastrophic data loss”.

Service Functions

* Consultancy
* Design & Build
* Provisioning
* Support & Maintenance
* Backup & Restore
* Capacity Management
* Reporting

….and so on.

Service Activities

* Consultancy
o Provide advice for storage technologies, products and architectures to support new business requirements and projects.
o Establish design standards and policies for storage
* Design & Build
o Define functional & non-functional requirements for new storage solutions
o Produce designs for new storage solutions
o Build new storage solutions
o Define end execute functional and non-functional testing for new storage solutions
* Provisioning
o Allocate new storage resources in response to service requests
o Provide storage network connectivity as required
o Decommission storage resources when servers are decommissioned
o Procure new or additional storage products as required
* Support & Maintenance
o Maintain all hardware, software and firmware revision levels within the storage infrastructure
o Provide support for storage hardware and software products and resolve issues
o Escalate issues to external service providers & vendors
* Backup & Restore
o Define and maintain backup schedules
o Perform ad-hoc restores in response to service requests
o Change tapes in tape silos and transport to and from offsite storage

….and so on.

Now decide who does what

Look at your existing teams, resources and skills and compare these with your service & activity catalogue. Where skills already exist within your organisation, can these individuals or teams take on additional storage-related activities? For example, can your Capacity Management team take on responsibility for storage in addition to distributed and mainframe servers? Can you networking team take on responsibility for Fibre Channel network infrastructure? Can Operations take on responsibility for running ad-hoc data restores? If these teams don’t have the skills or bandwidth to take on additional activities they may be able to do so with the addition of more staff and/or suitable training.

Alternatively, it may make more sense organisationally or simply be more cost-effective to set up a centralised group to manage every aspect of the storage service and carry out all the activities within the catalogue. It may even be possible to outsource the whole thing. In practice there are no hard and fast answers here and whether your organisation adopts a distributed, centralised or outsourced storage support & operational model – or combination thereof – will depend on many factors in addition to the ones mentioned here. The important thing is to define what your storage service is, what it’s made up of and to ensure that someone has ownership of and accountability for each aspect of it down to the activity level.

Time for some Processes

So far we’ve defined the “what” and the “who”. Now for the “how”, i.e. how do we ensure that each of the activities making up our service is performed to a suitable standard and, importantly, that it’s done in a consistent fashion? Performing any activity and achieving a defined end goal will typically require a number of people, who may be from different teams or in different locations to work together, each performing specific tasks all of which need to happen in a certain sequence.

This is where our processes come in. A process is simply a way of ensuring that what works well is remembered and can be reused. Processes help guarantee a repeatable and consistent outcome and save time by eliminating wheel-reinvention and also make it clear to all the parties involved what their responsibilities are. Here’s an example of a simple “straw man” process for ad hoc data restores: -

This is of course a very simple process, but this can be done for pretty much any process, even if it’s much more complex and involves many more teams in it’s execution (if you can’t draw a picture of a process it’s arguable that it’s too complex and needs to be broken down into sub-process, but that’s another story…).

Going back to our service and activity catalogue, a process diagram should be developed for each activity (e.g. “produce designs for new storage solutions”, “decommission storage resources when servers are decommissioned”, “escalate issues to external service providers & vendors” etc.) Once each process has been documented in this way, it’s clear to every group and individual involved in delivering the overall storage service what they need to do, in what order and, importantly, who they need to communicate with in order to expedite the activity in question. The chances of things being done in the same, optimal way each time are drastically increased and, hopefully, the margin for error is reduced meaning fewer change-related problems and outages.

There is another stage which may be required. This is to define the procedures to be followed within each process step. We’re not going to go into these here as technical procedures are inevitably highly dependant upon the technology being used, other than to say that it’s important that these are documented as well.

That’s Quite a lot to Think About

Absolutely. And remember, we’ve just been using some generic examples and simplified processes here. In reality, storage service and activity catalogues, processes and procedures are often highly complex, reflecting the nature of the technology itself and the extensive skillsets required to make it all work and manage its continuous development.

When you look at it in this level of detail, the “throw it over the fence and hope for the best” approach to Production handover is clearly a complete non-starter. Although it will add a significant amount of time and effort to your storage project, time spent on getting this stuff right up front will be rewarded many times over in the form of smoother implementation and less troublesome operation during the lifetime of the solution.

Selecting the right tools for the job

Another benefit of defining service, activities and processes is that it makes it much easier to select storage management tools or indeed the storage technologies themselves. The marketplace is awash with software tools which have an often bewildering level of extremely powerful functionality, but how can you possibly hope to decide which is the right one for you unless you work out what you need to do and how you want to do it?

If you are having difficulty deciding between vendor A and vendor B’s storage devices, it may be the case that only vendor A’s tools can be used in the way you want to deliver your storage service. Alternatively, if vendor B’s tools are half the price of the competition’s offerings, it may be worth subtly redesigning the way your processes work so that you can use the more cost-effective technology. But be careful not to design your processes around the technology’s functionality as you will inevitably lose sight of the real business requirements you are trying to satisfy in the first place.