Managing big data storage: the unified platform open-source approach
Red Hat outlines a unified, open-source approach to big data
Deployment of new and advanced technologies, support for an increasing number of devices, increasing virtualisation and the general expansion of business have dramatically altered the volume and nature of the workloads being handled by data storage systems.
The need to implement a strategy to deal with growing storage requirements is one that resonates with IT managers and CIOs across the region. According to Gartner, the Middle East and African IT infrastructure market, comprising servers, storage and networking equipment, is forecast to reach $3.9bn in 2013, which is a 4% increase from 2012.
In a race to accommodate the deluge of data, all too often enterprise IT teams are forced to react by creating storage silos, each with its own IT operations model. As data centres have been scaled up over time, all too often, these new storage systems have been added without sufficient consideration for what was previously deployed.
This has then led to a scenario wherein there exists a storage silo for each application workload: database data; a silo for shared file data; a silo for Web object data; and so on. This reactive approach can not only increase the capex for storage, but can create a huge impact on on-going operational expenses owing to the different management tools, different provisioning tools and different skill sets required. Given the size and rapid growth of data, and the prohibitive cost of copying large data sets around the enterprise, organisations simply cannot afford to build dedicated storage silos.
Virtualisation is now being widely employed by enterprises in order to both reduce infrastructure costs though better utilisation as well as to pave the way towards cloud deployments. Since the success of these virtualisation deployments depends on shared, network enabled storage infrastructure capable of eliminating the disparate silos associated with various applications and workloads, data storage and data management are now seen as a top priority for IT managers.
In addition, a centralised approach to data management is no longer feasible in the age of big data. Data sets are too large, WAN bandwidth is too limited, and the consequences of a single point of failure are too costly. A big data storage platform must be able to manage data through a single, unified pool distributed across the global enterprise. Instead, managing storage though a unified platform is a viable solution that not only simplifies data access and management but also greatly improves operational aspects such as power, cooling and space utilisation in data centres.