|
Myriad emerging technologies crowd an already full marketplace, yet most
companies are still trying to understand not how to integrate even another
technology but rather how to address the issues in the current operation.
While
it remains a good market for hiring people, Storage Administration is a
relatively new competence and not readily available. Where there are people,
they are expensive and unless there's a lot of on-going high-end work, they will
bore easily once the initial re-architecture is done. Also, the mix of
competencies necessary in a storage administrator encompasses technical
architecture, operations expertise, and business acumen. This is not an easy
mix to find, even in these days., The vast majority of consultants in the space
are affiliated with a given product company which means there's always a push to
buy product and also their focus is on integrating their product, not a
holistic, solution oriented approach. What is needed is independent consultants
not bound to sell hardware and software, with a focus on total solutions.
Classic
Case
Any space where a gap exists between what is known and
what needs to be known demonstrates a classic need for independent consultants,
which are organizations that leverage expert knowledge via defined methodologies
to provide predictably high results for their customers. The unusual thing
about the storage world is that the gap opened so fast and so wide. An industry
focused three years ago on how to double disk drive density, suddenly finds
itself trying to explain to Joe Admin the difference between switch and host
based virtualization. Good luck! Joe is still trying to figure out why the
backup system he bought 18 months ago still doesn't work and how he's going to
tell his boss.
In considering any large IT investment, four crucial elements
must be harmonized – Technology (features and functions), Cost, User
Requirements and IT Organizational Readiness. With margins falling on hardware
even faster than the stock value of the respective manufacturers, pressure on
Sales staff to "move iron" is intense. The recent overlap of features
that were once done only in hardware or only in software further confuses the
buyer and leads to analysis paralysis (ref. Joe and his virtualization lesson),
further delaying the sales cycle and exacerbating the intense pressure to cut
prices and close the deal. These factors put tremendous focus on the first two
(albeit, important) elements, Technology and Cost, but in many cases to the
detriment of the other two crucial elements. Without fully mapping the
technology to the end user requirements and thoroughly understanding the
implications on the IT organization in terms of
training,
competency development, policy changes, process revisions, etc., the best
technology at the lowest price does little to guarantee a successful outcome.
Historic Structures Become Today's Problems
The data
center historically was comprised of three groups – the systems group, the
network group, and the application group. In this world, the storage lived in
the system, voice and data services lived in the network, and the database lived
in the application. Then came the Internet, fiber, B2C, B2 – the attendant
history of explosive storage needs fueled by wide data pipes and rich media
applications, etc. is well documented. With the introduction of network based
storage, i.e. SANs, IT
delivered on the promise of scalable, manageable, centralized storage. The only
problem is someone forgot to tell the CIO that change has three elements that
need to be addressed – technology, process, and people. In the rush to
implement the technology, the focus on process and people was all but abandoned.
This convergence of systems with storage, networks and applications with
databases leads to three problems for the IT executive
1. The technology problem – the convergence is
driving a new set of technologies that will be needed to manage and support a
large, complex, heterogeneous storage environment. From storage resource
management (SRM) to virtualization to storage element management, there are over
250 venture financed
companies hoping to explain why theirs is the best product.
2. The business problem – with demands for storage
continuing to accelerate and now a heightened sensitivity to backup and restore
and disaster recovery, the only effective way to control, scale and manage
storage is in a centralized, network based configuration. Attempting to
continue to scale a direct attached, distributed model is doomed to fail. This
means that the business is at risk of being unable to expand unless there is a
cohesive storage strategy in place and an IT organization capable of rapid
response.
3. The organizational problem – the competencies
necessary to support a network based storage environment no longer reside in one
place. From the storage arrays which typically would belong in the systems
world to the fabric which typically would belong in the network world, the old
access controls – policies on who had which admin password, etc. – all
come up for debate. Leaving current organizational structures and policies in
place will handicap the ability to deliver quickly and leave IT viewed as inept
and stumbling over itself as it hands off tasks internally all the while failing
to meet the business needs.
Case Studies
A global manufacturing company
spent in excess of $2M on a collection of hardware and software to pursue a SAN
implementation to meet the growing volumes of data being generated by a number
of key applications. Six months later, the implementation was hopelessly
delayed, due predominantly to issues associated with making backups work in the
SAN environment. Needless to say, this was a technical issue related to the
fact that SAN based backups require knowledge of more than just the backup
software interface. What is interesting to examine is that while the problem
manifested as a technical issue, it could easily have been resolved had the
organizational issue been pre-empted; for example, had personnel with
insufficient competencies trained specifically for SAN management and backup,
prior to implementation. What is even more glaring is that this organizational
problem manifested as a technical issue, creating a business problem on two
fronts: the significant capital investment had not yet begun to show any ROI,
and the very problem for which the investment was made was now even more acute.
As such, the failure to develop appropriate competencies within the staff
resulted in technical difficulties which created negative business impact.
While this is not necessarily a new phenomenon, it is all too familiar to people
in today's storage world.
Another example is that of a Life Sciences
company which purchased a large storage solution to be integrated with its
existing Unix environment in a SAN configuration, to facilitate grow-on-demand
volumes for a mission-critical application. While everything looked OK on the
architecture diagrams, a lack of familiarity with the intricacies of HBAs
resulted in the company needing to do an OS upgrade on all their Unix servers in
order to get them on the SAN fabric, a small nuance that became clear only weeks
after the storage hardware arrived. Nowhere was this a greater inconvenience
that on the very servers that were deemed mission-critical, for which the SAN
was purchased to eliminate the need for downtimes. |
