HDD with
3D storage and parallel read/write |
. |
Since many
years I am searching reading and researching about digital storage devices used
in Computers like floppy- disks, HDD and CD/DVD etc. Specially in what format
data/information is stored on them and how data is read from them and how data
is written to them. |
I read mostly
about ram, floppy disks, cds/dvds, tapes and many other old devices. |
I was able to
understand some, some went above my head and I made some ideas myself. Specially
about format for storing data/information on HDD and also how to write
data/information to HDD and how to read data from HDD. |
. |
Ideas I made are as follows:- |
. |
1) |
How to store data on HDD: |
|
. |
Now a days data is stored in 2 dimensions on surface of HDD platter called tracks and sectors. Smallest unit HDD can read/write is called 'sector' and smallest unit within sector is bit which are actually written to or read from surface of HDD. 'Bit' can represent 2 states, 0 or 1 only meaning only 2 type of values can be stored in 'bit'. Where as my idea is to store entire byte at 'bit' location. Meaning instead of storing '2' type of values at bit location I want to store '2 power number-of-bits-in-byte' type of values in bit location. For example a byte of 8 bits there will be total 256 values. And of course now bit location will be called byte location and bits will be inside it in 3rd dimension |
This way HDD capacity will increase with size of byte.
|
. |
2) | How to do data read/write on HDD: |
|
. |
This I will explain with the following example and I will try to keep it easy. Think of a HDD of any standard size with only 4 tracks on it and only 4 sectors per tracks placed at equal distance. My idea is that there will be as many read/write heads as there are sectors on HDD. With each head stationed on start of some sector. Either heads or platter of HDD will move and other will be unmovable. So on any read/write request for a cluster, movement will be only for the length of sector. In other words read/write process for entire cluster will be done in parallel. Plus read/write of more then one cluster in parallel will be possible also. All this in movement of 1 sector only. |
Thus improving the HDD performance in term of time. |
. |
|
And if I further pack data meaning make bit location represent entire sector then there will be no need of moving platter or head as all head will be now stationed right above some sector which they will read or right without need of any movement. Not only it will increase HDD size (in multiple of sector size in bits) it will also increase speed of data read/write also. |
. |
|
In my system of things 'bit location' will be called something like ASL 'actual storage location'. Because actual storage will be done at these locations. |
. |
Further advantages: |
This type of HDD will provide many other interesting possibilities also. Like:- |
1) Entire HDD will be able to have it's own default sector and byte size. |
2) Each partition will be able to have it's own default sector and byte size. |
3) Each folder will be able to have it's own default sector and byte size. |
4) Each file will be able to have it's own sector and byte size. |
. |