Current Filter: Storage>>>>>Technology Focus>

   

WHAT ARE THE OPTIONS?
Traditionally, network-attached storage appliances (NAS) are used to further consolidate data. In each of these cases, data is organised into hierarchies within the file server or appliance, where some properties (or metadata) about the files stored are maintained. These cases also provide access to storage via a POSIX interface, most notably CIFS for Windows clients, and NFS for UNIX clients.

A newer option for unstructured data is object storage. The most popular example today is Amazon's Simple Storage Service (S3), which is accessible via a REST API interface that any client speaking HTTP can use directly. Mobile devices accessing Cloud applications were the first to use object storage for their documents and media files as these devices do not speak POSIX only HTTP. Any client or server in a data centre can also speak HTTP, though the bigger difference with object storage is that custom metadata is stored within the object vs. in a file system.

And yet, despite all these great benefits, object storage has taken an age to gain traction, mainly because of a lack of customer awareness and a killer app to drive its adoption.

Challenges of storing unstructured data:

1. File system Limits - Unstructured data is growing beyond the point file systems were designed to support. One example of a challenge is managing multiple petabytes in capacity, where single file systems cannot scale to tens of petabytes. Another challenge is managing tens of billions of files in a single file system. Whilst capacity or file counts can be distributed across multiple NAS systems, that in itself creates a management challenge to operate a sprawling number of files, and connecting the right applications to the right data they need on the right NAS system.

2. Consolidated Management - The proliferation of systems to store unstructured data creates a management burden where more administrator resources are needed. Taking cycles away from enterprise IT can relieve a cost burden, but can also lead to reaching the limit of what can be managed. Lines of business have grown shadow IT in their groups, which has lead to corporate data being stored in the public Cloud outside the control of IT.

3. Data Analytics - Data management and analytics capabilities cannot take advantage of metadata beyond the basic information that a file system stores. This is less a challenge and rather a lost opportunity as business leaders look to IT to enable better analytics and value from stored data.

4. Inefficient Utilisation - Storage for unstructured data has classically been difficult to expand and not feasible to procure in a pay-as-you-grow way. The norm has been to pay up front for the anticipated capacity that is needed over the useful life of a hardware solution, three to four years. In addition, enterprise IT also tends to hedge risk by overestimating the amount of storage that is needed, leading to many storage solutions failing to reach even 50 per cent utilisation over their used life. This is true whether using NAS or object storage solutions, but it is more prevalent in NAS.

5. Cost - Without factoring inefficient utilisation, the price/capacity ratio for conventional storage has not seen the same positive improvement that other areas of infrastructure like server/compute has. Enterprises cannot do more with the same without finding ways to improve price.

Page   1  2