Example Web Map
If you think about it, the World Wide Web is simply a massive storage repository that provides ubiquitous access to content over the Internet using HTTP. The Web browser is the primary user interface, which presents HTML pages containing hyperlinks to other HTML pages and specific content types including text, documents, images, video, etc. The hyperlink associated with a URL (uniform resource locator), enables the user to navigate to and access another piece of content.
There’s a lot of flexibility to name the resource/content within an HTML page such as:
“…there is an interesting item you should read to learn more about the subject.”
The resource name or label is “interesting item”, which could be a PDF document, an image or another page. The hyperlink’s URL is the address to access the file and below is the full URL for “interesting information.”
http://www.caringo.com/downloads/whitepapers/caringo_whitepaper_newmillennium.pdf
It is similar to the paradigm of the file system where you assign a file name, MyFile.doc, but store it in a specific location. In order to access the file you need to know the path to get to it.
C:\MyFolder\MySubFolder’\MyFile.doc
Instead of having to traverse a folder\sub-folder hierarchy to open a particular file, one simply clicks on a hyperlink to retrieve content when surfing the Web. In this scenario the Web is like an infinitely vast, flat address space where all content, as perceived by the end user, exists at the same level (see Web Map image above). Sure, on the Web server itself the resource may be stored in a file system hierarchy, but that has no impact on the individual executing a hyperlink. The Web repository is huge and grows transparently to meet demand as it comes.
What is intriguing about this is how it relates to the buzz-word du jour, “cloud storage.” While a formal definition of cloud storage remains elusive, there does appear to be agreement that it enables access to storage resources via Internet protocol (public/private) that it is resilient, affordable, massively scalable and simple to use/manage. As a concept, it makes great sense, but where the rubber meets the road is determining the underlying technology that’s going to deliver on the promise.
Applications of all types have moved to embrace the Web to simplify their offerings and why shouldn’t storage. What we’re seeing today is the emergence of object-based storage with an HTTP, REST and/or SOAP interface that behaves somewhat like the Web itself. It is also the technology best suited to provide the underlying infrastructure for cloud storage. Several vendors big and small have introduced products/technologies into this new category of storage.
As an example, a file can be named or labeled, and stored through a content management system or file system. When ingested to object storage the file name becomes a part of the object metadata and stored with the file data. The address for retrieving that file is simply a URL linked to the file name or label for access and never changes.
One method is to link the specific file name MyFile.jpg or you can create a label as a hyperlink in a web page such as “Follow this link to my file that you’ll find quite interesting.” Below is the URL for both of these links and the content is accessed on a storage cluster directly connected to the Web. It’s simply raw object storage available on the Internet or maybe more appropriately, Web storage.
http://cas.caringo.com/760C9DFE1E6F5CE01AC75F715B9731CD
The cluster: standard x86 servers, massively scalable, highly resilient, self-healing, self-managing, with automated intelligent content distribution.
Sounds to me like all you need for “cloud storage” infrastructure.
0 Responses to “www.massive.con:tent/storage/repository”