Managing a Data Center

The CIO of a leading and rapidly expanding biotech company called his IT
staff and said, “We need to consolidate things into one physical location. We
must try to put them into a much smaller footprint.”

That is where the challenge began! In the past, the organization typically
dedicated one physical server to each application or user group. Today, the
organization has 3,000 research analysts who run multiple applications in their
areas of expertise on multiple physical systems. So now, the IT team is looking
for a solution that will allow them to operate more efficiently and at the same
time provide them with the flexibility to adapt and grow.

Since the researchers use the highly complex data modeling software, the
amount of data generated runs into over 70 terabytes. This data is critical as
it holds all the research findings. The future of organization depends on this
data. To add to their woes, the storage environment is highly heterogeneous in
nature with storage boxes from 2-3 different vendors. Managing it is a complex
task.

The IT team faces a dilemma while exploring various solutions to make their
100 servers and their storage environment more manageable. Should they simply
consolidate or is virtualization a better option?

“During the past 10 years we have experienced a lot of temporary growth. To
achieve better growth in future we are now restructuring our server and storage
environment,” says the IT Head. “We now think more than ever about
integration and simplification.” What approach should they adopt?

Dataquest spoke to two leading experts on what approach should the company
adopt in this current scenario. Their opinion reflects a system integrator's
and vendor's perspective to this problem case.

This question deals with complexity, one of the biggest and most common
problems IT managers are faced with. Managing a data center, today, can be a
nightmare, as the networks often include thousands of disparate appliances, each
requiring its own management console. As a result there has been a demand for
consolidated appliances, which offer multiple services, all from one device.

Many devices available in the market, today consolidate functions such as
firewalls, VPNs, intrusion prevention or detection services (IPS/IDS), and other
services into one appliance. This reduces the cost of purchasing multiple
appliances (CAPEX) as well as the cost of configuring and managing them (OPEX).

Besides, consolidated devices also offer the benefit of scalability. But this
scalability is possible only to a limited extent.
If you are looking out for a new firewall and know that six months down
the road you are also going to need VPNs, then purchasing a consolidated device
that includes both would be beneficial. This would enable the user to start the
VPN service, whenever he is ready to upgrade. The limitations, however, are that
one can only use a limited amount of services before needing to purchase
additional consolidated appliances.

There are inherent risks involved with the consolidated devices, alongwith
the benefits. One such risk to consider is the introduction of a single point of
failure in your network. Say you are using a consolidated device for firewalls,
VPNs, and intrusion prevention and the device fails. You would immediately lose
all connectivity. Depending on how separate appliances were deployed, you might
have been able to escape with just a partial failure if the services were
separate. However, traffic would still have to flow through the failed
appliance. In order to avoid this, it is often best to deploy appliances,
whether consolidated or not, in pairs to provide a backup in case of device
failure.

Another issue to consider is the level of management consolidation that
complements the feature consolidation. If several functions are consolidated
onto a single device but each function still requires its own management
interface, then you will still face the same problem of managing multiple
consoles. The problem is same as managing multiple appliances.

It is also important to consider the difference between consolidated devices
and virtualized devices. One needs virtualized solution along with a
consolidated one if he goes for more capacity and functionality in the future.
With consolidation, one would still need to deploy more integrated appliances
throughout the network.

Whether consolidated or virtualized, one would ensure that the product has
high availability features to maximize uptime, as well as appropriate management
consolidation features to accompany the consolidation of functions.

If one decides to go virtual, he must keep in mind that virtualization
technology is still evolving. One should deploy this technology internally or
via on-demand services incrementally so he can monitor initial deployments,
closely. This will enable him to validate that virtualization approach can
clearly improve the performance of existing systems without becoming a single
point of failure or disrupting any operations. Virtualization simplifies
management while increasing scalability, but it is wise to take an incremental
approach in order to ensure the success of your deployment. 

The problem presented here is a classic case of conflicting objectives that
every company wishes to achieve: reduce footprint and at the same time provide
flexibility to grow rapidly. Consolidation and virtualization are two important
technologies that help to simplify IT infrastructure. An interleaved
architecture enables the corporation to achieve greater levels of manageability,
business continuance and also helps in managing the information throughout its
lifecycle (ILM).

An important point to note is that virtualization and consolidation are not
mutually exclusive options. Both of them can bring rich dividends in terms of
lower TCO, reduced costs, better management capabilities and increased
utilization when strategically deployed after careful planning.

The answer here is that they should do both. The CIO's concern for
footprint reduction clearly calls for consolidation.
That is based on a business need.  However,
there are other facts to be considered as well. First, there is an existing
heterogeneous storage that contains critical data. Second, there is a need to
protect investment from hardware perspective. Third, the company already has 70
TB of useable data and expects substantial growth in years to come. And, last
but not the least, the issue of management complexity looms large. All these
factors point towards the necessity to virtualize storage.

Hence, I would recommend that at the first available opportunity the company
should plan and implement virtualization. This will bring all the existing
storage devices under the virtualization fold, thereby pooling the storage
capacities from across all of the storage devices. This will reduce the
management overheads as well.

As a second step, they can start consolidating storage by moving all the
critical data into fewer storage devices. Virtualization will enable this
movement without causing any application downtimes. This will also help
constructing a hierarchical storage structure by re-purposing the old storage
(protection of investment). At the end of the exercise, they will achieve the
consolidation and virtualization.

Why Should They Virtualize?

Effective virtualization of IT resources can help mask the complexity of the
storage networks. This in turn increases resource utilization, improves
personnel productivity and enables a tiered storage environment in which one can
match the cost of storage to the value of data. For a biotech company, the
propensity to accumulate lot of data is high.

Why Should They Consolidate?

Consolidating the storage environment should be a logical step in this
company's efforts toward simplifying the infrastructure. Consolidated
environments can be more resilient and more cost-effective to manage, reducing
planned downtime and TCO.

As Told to Shipra Arora

shipraa@cybermedia.co.in