Getting Real About Cloud Replication

With vendors consistently selling the idea of inexpensive cloud storage, there is a trend towards IT administrators wanting to backup every change as it happens and keep it for ever. Seemingly not a day goes by without some over zealous marketeer making claims that defy the laws of physics.

Over the years we have worked with each new iteration of software that claims to be the new panacea to the problem of replicating large volumes of ever changing data over a WAN and storing it indefinitely.

Whilst this will certainly sell more storage it encourages sloppy system administration practices and increasingly unrealistic expectations from end users, which the now overloaded infrastructures struggle to meet.

Rather than approach every application in an infrastructure as mission critical, a huge amount of time and money can be saved by assessing the actual service level expectations of your departments and users.

We get asked all the time to keep data on the platform for 7 years for “compliance reasons” often with a very loose justification. Whilst it would be a simple thing to take this on and worry about it in 7 years time, far better to understand the real world requirements.

In real world examples such as the Financial Services Authority requirements, they may make a request for the information, either for evidence or as part of an investigation but there is a very generous time-frame given for the data to be supplied.

Given the speed at which these investigations move, its not unreasonable to expect a number of weeks in which to satisfy these requests.

We always encourage some level of interaction with stakeholders, who, once aware of the potential cost implications, usually accept a reasonable internal SLA which in turn allows this information to be stored on archival storage for a far lower cost per GB.

There are highly efficient and intelligent software solutions that will readily manage your data types and apply your given service levels however the actual definition of the service level for each type of information must be approached by IT and the business as a whole to avoid spiraling data costs.

By injecting a little realism and keeping our feet on the ground while our heads are ever increasingly in the clouds it’s possible to use the most appropriate cloud solutions for replication, backup and archive of the specific data type. The proliferation of API driven Object storage gateways meant that managing multiple Cloud services is becoming simpler by the day.

As ever this means the demands placed upon the IT professional to up skill and maintain their continuing professional development must not be overlooked by Businesses that want to remain current and competitive.