Subscribe

Cloud recovery challenges

Companies are under pressure to ensure data is securely stored and backed up.

Martin May
By Martin May, Regional director (Africa) of Extreme Networks.
Johannesburg, 01 Aug 2013

Data backup, storage and recovery processes are fundamental services in almost every business. Recent changes in the IT environment, notably the arrival of cloud computing technologies, are responsible for significant advances in traditional data management patterns.

Representing one of the most important shifts in computing today, cloud technologies provide greater levels of provisioning and automation, while helping organisations become more flexible and responsive to change.

The cloud has also made a significant impact on companies' ability to better allocate their computing resources in order to improve application performances, boost efficiencies and reduce operating costs.

Companies are under increasing pressure to ensure data backups are performed efficiently and sensitive business data is held in a secure storage repository. Recovery has to be effected faultlessly and reliably against elevated benchmarks for speed, accuracy and completeness.

Backup goals

In order to achieve these goals, organisations have been driven beyond traditional product-centric or device-centric approaches. Today, evolving cloud-based, service-centric data backup, storage and recovery models are gaining in popularity on a global scale.

In this light, an increasing number of companies are opting for fully outsourced data services from the many backup-as-a-service (BaaS) providers mushrooming in the global marketplace.

BaaS providers give their customers the opportunity to delegate the often arduous responsibilities associated with maintaining ever-expanding in-house data storage repositories and running (and updating) increasingly complex management solutions, freeing up highly skilled IT staff to work on other projects.

BaaS providers deploy cloud-based backup strategies that retrieve data from their customers' servers, databases and other systems and store it in a virtual repository. From the users' perspective, these solutions are able to simplify IT infrastructures, limit management overhead and improve responses to changing organisational demands.

Service providers are continually innovating increasingly agile, scalable, fail-proof solutions. Examples include hybrid-cloud solutions that manage both cloud-based and traditional on-premises data.

By allowing companies to provide and manage some of their resources in-house (in their own data centres, for example), those with concerns over the precise location of key business data can be placated.

That said, hybrid cloud solutions are far from a compromise. They give users the flexibility to store certain data at traditional locations and other data in a shared cloud repository - and the ability to change the model at will as business needs evolve.

Because hybrid solutions also enable significantly higher levels of automation, orchestration, provisioning and deployment in traditional processes, they have implications for the streamlining of server, storage infrastructure and data centre designs going forward.

First time lucky

An important role of the BaaS provider is to reduce backup times by taking advantage of the most efficient use of bandwidth and an optimised storage footprint in which, ideally, the entire dataset is captured on the first backup, with only changed blocks captured on subsequent backups.

One of the challenges associated with implementing a hybrid cloud backup/storage/recovery system is that legacy data centres are frequently highly siloed environments incorporating islands of servers, storage devices, networks and applications.

Those with concerns over the precise location of key business data can be placated.

Therefore, many steps are required in order to provision the infrastructure for a new cloud application, including planning for and installing new equipment, configuring network and storage, loading operating systems - which need to be co-ordinated across virtual machines (VMs) - and the setting up of control processes together with the policies associated with all storage resources.

The efficiency of the solutions selected and their linking methods will have to meet a range of demands for efficiency and control, one of the most important of which is latency limitation.

Latency, the measure of time delay experienced in a system and often linked to bandwidth availability, could be impacted by a number of decisions, including the number of router hops in the users' network (or the BaaS provider's own network).

For businesses looking to boost end-to-end quality of service and maintain customer satisfaction, every millisecond counts. For this reason, latency has been viewed as one of the barriers to broad-based hybrid cloud adoption. It is said that a half-second delay will cause a 20% drop in Google's traffic, and a tenth of a second delay could cause Amazon's sales to fall by a percentage point.

While latency in itself is not a major issue, it is its unpredictability that is at the heart of the problem. By using dedicated connectivity links to a BaaS provider - versus the Internet - or linking via carrier-neutral co-location data centres and connectivity providers, the uncertainties around latency can be minimised.

Although there are several complicating factors, careful planning in conjunction with a specialist vendor or reseller can ensure that bandwidth and latency pitfalls are avoided and applications run smoothly in the cloud.

Share