< ? php //If there is analytic campaign data, attempt to get the campaign_guid from that cookie if ( 1 === preg_match( '/pk10mkto-([0-9]+)/', $_COOKIE[ '__utmz' ], $match ) ) { $campaign_guid = $match[ 1 ]; } ?>

Object Storage: If Resilience Matters, Location Matters

Yellow, red, green and black pushpins in a map
Shares

If you read our previous blog on object storage , Peak 10 Object Storage in particular, you know that one of its key attributes is resiliency. How Peak 10 helps ensure that resiliency — as well as data integrity — is among the many features that distinguishes it from the competition. It’s largely a matter of location.

Data Protection via Replication

Generally speaking, object storage protects data using one of two methods: replication or erasure coding. Both are designed to address problems such as drive or array failures.

With replication, each object is fully copied to a specific number of nodes. At least two data copies are stored although, depending on the vendor, the number of copies and how and where they will be distributed may vary. The idea is that if the original object becomes damaged or lost, an exact copy will still be available.

Among the issues with replication, however, is that multiple copies of data take up storage capacity — a lot of storage capacity if the data files are large and spread across different data centers. Many vendors charge for storage at each site when multiple locations are used, as well as for transferring data between the facilities. That can substantially raise the cost.

Data Protection via Erasure Coding

The other common data protection method is erasure coding, a parity-based scheme similar to RAID. Data is broken into fragments that are expanded and encoded with a number of redundant pieces of data. The data fragments are then stored across different storage nodes. If the data becomes corrupted, it can be reconstructed using information about the data stored elsewhere.

Erasure coding can be implemented locally or distributed, which means that fragments can be spread over multiple data centers, and the system can survive the failure of a full data center. Distributed erasure coding drastically reduces the extra storage needed to protect the data.

As with replication, cost can be an issue. Rebuilds require data transfer between data centers. If the vendor is charging fees for transferring data between data centers, the price goes up.

Peak 10 Geo-diverse Data Protection

Peak 10 employs a distinctive approach to object storage data protection by combining erasure coding and replication, providing customers with the “best of both worlds.”  This approach optimizes efficiency and cost. Data is erasure coded and then automatically copied across three different ISO/IEC 27001:2013 certified data centers.

Many vendors make copying data across multiple data centers an “option,” one that customers must pay for and are required to manage. While it’s acceptable to copy data within a single data center, doing so only makes the data fault tolerant. If something happens to the original, there’s still a copy. The problem is that if the data center goes down or becomes inaccessible for any reason, both the original and copy are unavailable.

A good analogy is a set of car keys. You likely have a duplicate set of keys for your car. If you lost one set, you have another set as backup. However, if you had both your primary and duplicate keys on the same keychain and lost it, you’d be out of luck.

Peak 10 not only erasure codes data to three separate data centers. It erasure codes data to three geographically diverse — geo-diverse — data centers located at least 200 miles apart. In addition, the facilities are positioned in a triangular configuration relative to each other. If a region-specific disaster took out one facility, the other facilities would likely remain safe. This significantly reduces the chance of ever losing both the original and its copies, and enables Peak 10 to deliver maximum availability and resiliency. Not surprisingly, Peak 10 is able to offer SLA-backed 99.999% availability.

Peak 10’s approach also negates the cost issues associated with replication and erasure coding. The base price for Peak 10 Object Storage covers storage at three data centers. There’s no additional storage charge for each data center, and there are no fees for transferring data between data centers. Nor does Peak 10 charge for API transactions, and the 24/7/365 technical support that is included. Taking all those factors into consideration, the total charged by many leading object storage vendors for what Peak 10 provides as standard could easily run two to three times as much.

There’s More

Of course, there are many other distinctive features of Peak 10 Object Storage, from its anytime, anywhere access over the Internet via S3-compatiable API to its compatibility with Commvault® applications and NetApp AltaVault devices for backup. Watch for future blogs in which we discuss these differentiators.

Fine tune your content search

About Peak 10

"Our values are the foundation for everything we do at Peak 10, and are ultimately what enable us to earn our customers' business and their trust."
David H. Jones,
Board Member, Peak 10 + ViaWest