Massive Scalability Requires More Than Just a Massive Storage System

In a previous blog entry, I took a look at some of the factors driving the need for a scalable architecture that meets today’s enterprise archiving needs. In this blog entry, I’ll take a closer look at how Permabit‘s Enterprise Archive delivers on some of these requirements.
In the past, and even to a large extent today, the idea of building a massively scalable storage unit for your archived records was pretty simple to understand – build a bigger building or knock out a wall and put on an addition. Even today companies too often equate archiving with keeping more boxes of paper and tape in ever bigger buildings with your records management provider.
However as companies move towards archiving data on disk-based storage systems, you can’t just always build bigger buildings or knock down walls. If anything, companies want to store more data in a smaller footprint. Making it more complicated, companies are creating exponentially more data than they were 10, 5 and even 2 years ago and keeping it for longer periods of time. Factor in mobile devices that manipulate existing data and create new data and the increasing use of video in corporations and the result is millions, billions and even trillions of file-based data elements that create thousands of terabytes of data.
In the past disk-based storage systems contributed to the problem since they were designed to support daily operations, not retain data for long periods of time. This limitation forced companies to take great care in selecting a platform that scales to meet their data archiving and storage capacity needs short and long term. Using disk-based storage systems are becoming the preferred platform for storing archived data, it may add to the difficulty in selecting a system since the decision has more far-reaching implications.
Specific concerns that a company may have about using a disk-based storage system for archiving include:

  • Cost
  • Data migrations
  • A single disk’s finite storage capacity
  • Implementing new disk storage technologies as they become available
  • Storage system upgrades
  • System management and efficiency

Permabit‘s Enterprise Archive addresses these specific reservations that companies may have around using disk to store their archived data since “massively scalable” no longer means bigger buildings but better data reduction and data retention technology in the same size footprint. Permabit’s Enterprise Archive uses a grid storage architecture that facilitates easy storage capacity expansion and technology refresh as well as supports transparent advances in storage system technology (faster processors, higher speed controllers, faster network interconnects, larger capacity disk drives, deduplication, etc.) in its design.
Enterprise Archive’s grid storage architecture uses nodes that are based on off-the-shelf hardware components (servers and disk drives). Companies can then upgrade to newer, faster processor and storage technologies simply by introducing new nodes into the Permabit Enterprise Archive configuration. The process is seamless without user involvement nor interruption. Once introduced, data is then automatically redistributed from existing nodes to new nodes or, assuming companies want to replace an existing node with a new node, all of the data is then migrated to the new node. This configuration eliminates the normal management hassles, resource impact and costs (internal and external) of data migrations that technology refreshes introduce while providing a transparent upgrade path to new storage system hardware technologies.
Enterprise Archive’s grid storage architecture is based on Permabit’s RAIN-EC technology helps to ratchet up its maximum storage capacity. The Data Centerseries of Permabit’s Enterprise Archive scales to 96 TB in individual grids (up to 32 grids) which can then be pooled together for a total system capacity of 3 PB. However when one starts to think in terms of archiving data for 10, 30 or 50 years or longer, 3 PB may still seem insufficient. But then again, there are increased drive capacities on the horizon.
That’s where another component of Permabit’s Enterprise Archive feature set comes into place: Scalable Data Reduction (SDR). SDR performs data deduplication on data as it is ingested by the Permabit Enterprise Archive by comparing incoming segments of data with existing data segments. Storing like segments eliminates redundant data on the Permabit Enterprise Archive and provides companies with a higher net effective storage capacity than just the raw storage provided by the storage nodes.
SDR also drives down costs since existing storage capacity is used more effectively. Further, by coupling SDR with Permabit Enterprise Archive’s grid storage capacity, larger capacity drives can be regularly introduced that increases Enterprise Archive’s raw storage capacity capabilities which further contributes to keeping storage costs at a minimum.
In these respects, Permabit’s Enterprise Archive gives companies a means to massively scale a disk-based storage system without needing a massive building in which to store their archived data. By combining a grid architecture with its SDR deduplication and Permabit’s RAIN-EC technology (subject of another blog entry soon) , Enterprise Archive delivers the inherent attributes of infinite capacity, simplified management and long life that storage systems intended for managing archived data requires.
However disk-based systems designed for archive also need to address other concerns such as preserving the integrity of the data (digital fingerprinting) to withstand the possibility of multiple simultaneous hardware failures as well as providing content certificates to ensure the data satisfies legal audits. I’ll take a closer look at those features in an upcoming blog entry.


Click Here to Signup for the DCIG Newsletter!


DCIG Newsletter Signup

Thank you for your interest in DCIG research and analysis.

Please sign up for the free DCIG Newsletter to have new analysis delivered to your inbox each week.