Tiered storage in the cloud

BANGALORE, INDIA: Tiered storage, as we all know, is a hierarchical architecture of storage systems based on a set of parameters relevant to the organization.

These parameters could include data confidentiality, security, speed of back-up and recovery thus providing business continuity and compliance mandates which include duration of retention, all while optimizing the cost of the storage.

In today’s world though, data back-up is more subjective whereas traditionally, it was more objective. Simple parameters such as frequency of access or size of file are no longer decision-making criteria to determine the criticality of data.

ILM (Information Lifecycle Management), if implemented thoughtfully, is a probable solution to the above problem. ILM can be thought of as a rules manager that delivers the business logic to determine what storage type is the right fit for the data.

By applying policies that define the tiers of service and setting the rules on data classification based on set parameters that define information, data, and security service requirements such as retention, placement, protection, business continuity, archive, and compliance policies to information based on its value or business requirements, ILM practices can be optimized to deliver cost and time benefits.

In terms of business and technological value, there is a great synergy between cloud and tiered storage. Cloud storage is a great way of optimizing an organization’s op-ex and cap-ex, the former by reduction in the manpower, overall power consumption, space consumption etc. and the latter by lesser investments in hardware.

However, there have been a few inhibitors in quick and early adoption of cloud storage. Of this, the concern over the security of data residing in the cloud vis-à-vis in an enterprise data center has not gone away yet.

Being a young concept, the cloud has yet to establish itself as a secure environment. Also, the loss of visibility of the actual location of the data in the cloud is a concept that most data managers have not yet come to terms with.

Though theoretically, cloud storage is supposed to be unlimited, no one has really been tested on this claim as yet; hence, there is an apprehension on this claim and the scalability is still a question mark.

{#PageBreak#}

Since the cloud has not really been tested for its capacity and durability, the impact of cloud outages is still unknown and organizations are unsure whether the data in the cloud will be compliant with the business continuity SLA’s.

Hence, for the same reasons that the cloud acceptance has been low, it may not be a fair expectation to anticipate a change of behaviour in the acceptance overnight.

The solution might be in the form of a go-between. Just as ILM works in a traditional (non-cloud) environment, it can be implemented to tier storage in a hybrid (private and public cloud) environment.

Nowadays, readymade solutions are available for the purposes of data classification in order to get the right data at the right place in the tiered environment. Implementing such a solution can provide an organization the best of both worlds - optimum usage of the investments already made in enterprise storage by using it for the grade 1 critical data.

Grade 2 data can be initially located on the enterprise NAS or DAS, but can be further sifted depending on its behaviour in a live environment and optimized between the cloud and the enterprise storage.

The lowest critical data can be on the cloud from the beginning and would help the data managers understand the behaviour of the cloud thus helping make the decision on movement of the remainder of the data.

The author is Senior Vice President, Enterprise Technology at Symphony Services.