In our days intensive development is done to replace the hard disks used in computers with non-mechanical storage devices. A variety of SSD (Solid State Drive) technology exists, the most important are the flash and DRAM based solutions.
Common feature is the much higher transfer speed than traditional hard disk drives. Sequential reading and writing is significantly faster, but the greatest improvement is the random access time. While working with smaller files they perform much better than hard drives. Of course, SSDs are much more reliable from mechanical point of view, due to the lack of moving parts. Their disadvantage is the still higher price and lower capacity. The flash-based SSDs have limited number of write cycles, which is not a problem in common use.
The topic of my thesis is the development of a new SSD device and its major components, built from SD cards. The lower write speed of the cards can be increased using more cards and RAID organization. Thus redundancy can be built in as well, which allows replacing damaged or worn out (during too much write cycles) cards without losing data. Using many cards a relatively large capacity is achievable, which can be competitive with hard disks. The device will be connected to a computer on ATA or SATA bus and it will be used as a traditional hard drive.
My duties were to get to know the specification of the SD cards and the different RAID versions, then to examine and evaluate the RAID possibilities with the cards in terms of speed. After that I dealt with the development of the SD card handler hardware. In my thesis I present the types, characteristics, advantages and disadvantages of SSDs, then I discuss the technologies I have used. After that I describe the RAID configuration options, then the major steps of development work, the resulting hardware’s structure and functionality. Finally, I discuss the possibilities for further development and the additional steps required to complete my work.