The Ascent of Object Storage

SOA is Dead; Long Live Services

By G C Network | January 7, 2009

Blogger: Anne Thomas ManesObituary: SOA“SOA met its demise on January 1, 2009, when it was wiped out by the catastrophic impact of the economic recession. SOA is survived by its…

2009 – The Year of Cloud Computing!

By G C Network | January 6, 2009

Yes, everyone is making this bold statement. In his article, David Fredh laid out the reasons quite well: The technological hype has started already but the commercial breakthrough will come…

Salesforce.com and Google expand their alliance

By G C Network | January 5, 2009

In a Jan. 3rd announcement, Salesforce.com announced an expansion of its global strategic alliance with Google. In announcing the availability of Force.com for Google App Engine™, the team has connected…

December NCOIC Plenary Presentations

By G C Network | December 31, 2008

Presentations from the NCOIC Cloud Computing sessions held earlier this month have been posted on-line in the Federal Cloud Computing wiki. The event featured speakers from IBM, Cisco, Microsoft, HP,…

Booz|Allen|Hamilton Launches “Government Cloud Computing Community”

By G C Network | December 30, 2008

As a follow-up to a Washington, DC Executive Summit event, BoozAllenHamilton recently launched an on-line government cloud computing collaboration environment. In an effort to expand the current dialog around government…

Is Google Losing Document?

By G C Network | December 29, 2008

John Dvorak posted this question on his blog Saturday and as of Sunday evening had 52 responses! This is not a good thing for building confidence in cloud computing. Or…

Cryptographic Data Splitting? What’s that?

By G C Network | December 26, 2008

Cryptographic data splitting is a new approach to securing information. This process encrypts data and then uses random or deterministic distribution to multiple shares. this distribution can also include fault…

Now really. Should the Obama administration use cloud computing?

By G C Network | December 23, 2008

It’s amazing what a little radio time will do! Since Sunday’s broadcast, I’ve been asked numerous times about my real answer to the question “Will ‘Cloud Computing’ Work In White…

NPR “All Things Considered” considers Government Cloud Computing

By G C Network | December 21, 2008

My personal thanks to Andrea Seabrook, Petra Mayer and National Public Radio for their report “Will ‘Cloud Computing’ Work In White House?” on today’s “All Things Considered”. When I started this blog…

HP Brings EDS Division into it’s cloud plans

By G C Network | December 18, 2008

The Street reported earlier this week that Hewlett Packard’s EDS division has won a $111 million contract with the Department of Defense (DoD) that could eventually support the U.S. military’s…

Over the past few years, the data storage market has changed radically. The traditional hierarchy of directories, sub-directories, and files referred to as file storage has given way to object storage, individual storage objects. While file storage was designed to help humans interact with data, object storage is all about automated efficiency.
User expectation of data usage drives file storage repository design.  In this structured data model, all folders and names are organized to support a pre-defined business process or model. The file system also associates a limited amount of metadata (i.e., file name, creation date, creator, file type) with the saved file. Finding individual files is done either manually or programmatically by working through the hierarchy. The file storage approach works well with data collections but can become very cumbersome as data volume grows.
Object storage, on the other hand, is optimized for an unstructured data model. While this approach is not “human-friendly” it also doesn’t require prior knowledge or expectations of data use. Files are stored as objects in various locations with a unique identifier and a significant amount of metadata. The size of the accompanying metadata can range from kilobytes to gigabytes and often includes a content summary, keywords, key points, comments, locations of associated objects, data protection policies, security, access, geographic locations and more. Enhanced metadata enables a lower level of granularity when protecting, manipulate, and managing stored objects.
Specific business, technology, and economic drivers caused this significant market change. Business drivers include:
  • Rapid growth in amount and importance of unstructured data
  • Need to implement faster data retrieval based on identifying details incorporated in metadata that the operating system reads.
  • The requirementto apply organization to unstructured data resource through the use of text analytics, auto-categorization, and auto-tagging.
  • Increased legal and regulatory requirements for scalable data archiving and e-discovery
  • Enhanced business process and model flexibility enabled by the use of a flat storage structure.
From a technical point of view, object storage is far superior to file systems. This advantage is primarily due to its unlimited scalability and ability to be managed programmatically. It also:
  • Has fewer limitations when compared to the traditional file or block-based systems because of the flat data environment
  • Ability to customize metadata through arbitrary use of any number of data attributes
  • Global accessibility using HTTP(S) and REST APIs
From an economic point of view, object storage is also more cost-effective than file storage solutions, especially when storing large amounts of data. Since object storage solutions efficiently leverage unlimited scalability, organizations find that it is less costly to store all of their data. This advantage also exists in private cloud implementations where costs can be even lower than that provided by public cloud providers. Object storage is also much more durable than file-based alternatives.
The marketplace offers plenty of alternatives when object storage is the right answer. Access protocols, technology, and cost, however, varies widely. As shown in Table 1, storage cost for 1 terabyte of data for one year ranges from a high of $4,300.80 with data striping from QualityTech/QTS to a low of $47.88 from Wasabithat uses a more advanced erasure codingapproach. While location differences cause some cost variation, most of the variation can be attributed to design architecture and underlying storage technology.
 
Although this market survey is not exhaustive by any means, it highlights the importance of being an educated consumer when considering object storage solutions.  Other solutions aspects worth investigating include:
  • Complexity and performance across provider storage service tiers
  • Data immutability and durability
  • Speed of internal consistency across multiple copies of your data
  • Elapsed time to the delivery of the first byte of requested data
  • Use of active integrity checking
By all objective accounts, object storage is the right storage for large segments of an organization’s data holdings. This reality should lead to more effective due diligence and care when considering your enterprise’s next storage upgrade.


( This content is being syndicated through multiple channels. The opinions expressed are solely those of the author and do not represent the views of GovCloud Network, GovCloud Network Partners or any other corporation or organization.)

( Thank you. If you enjoyed this article, get free updates by email or RSS – © Copyright Kevin L. Jackson 2018)

Follow me at https://Twitter.com/Kevin_Jackson
Posted in

G C Network