The Storage Network

Storage Space Isn´t Everything

Probably nothing has increased so much over recent decades as cheap, readily available storage space. It is not so very long ago that people thought in megabytes of storage space – gigabytes, let alone terabytes, were unimaginable things of the distant future, reserved for the realms of supercomputers. Given the considerable amount of data which can be saved on even the smallest modern home or office PC today, one is therefore led to believe that all is taken care of on the data-storage front. Well, no. Availability and data-access times – not sheer size of storage space – are the true limiting factors in data traffic. Nonetheless, the quality of access times is often underestimated – especially when dealing with comprehensive web applications. Script languages such as PHP, for instance, often require up to a hundred tiny files in order to generate the content of a webpage.

An off-the-shelf hard drive today requires approximately 12 ms to reach the position of the desired data (average seek time). This means that it would take over a second to merely find the 100 files that have to be processed for the content of a webpage to be displayed. Now add to this the actual read process, the processing, the database access, and the data transferral to the webpage visitor. These individual factors quickly add up to more than 3 seconds of waiting time. No one likes to wait – and Internet users in particular do not like waiting for the content of a webpage to be displayed on their screens.

Apart from access times, the other important criteria that must be considered when dealing with data transferral are the availability of data and its secure storage. RAID solutions, which protect against the failure of individual hard drives, are, unfortunately, not a cure-all for the problems that go hand in hand with data storage. The failure of one hard drive generally causes a loss of performance which, when dealing with thousands of simultaneous page-views, can often not be compensated. A failure of the RAID system itself would paralyze the entire website. There are several cases where problems caused by such a total blackout have cost companies valuable customers.

In order to eradicate this weak point, IDNT saves all data on a storage network. This network consists of several stand-alone storage systems (storage nodes). Each storage system in a ganged pair contains identical data. This distribution of data also means that it can be read from several different sources simultaneously – thus reducing access times considerably. Individual, frequently requested files are actually stored on and accessed via the internal memory. This once again speeds up the already formidable access times of such a system.