Backup for eternity (2)

Which storage media is best for long-term archiving? This week: HDD, SSD and Flash-based USB!

As we have mentioned in our previous blog article of this series, enterprises are currently facing severe difficulties on how to store data. Big data has reached even the smallest company and more and more data is generated by online portals or IoT devices alike. While not all of this data will be necessary to store or archived for a long time, some – and more than in previous times because of the overall increase – has to be archived because of new or existing legal or regulatory frameworks.

But were to store all this additional data? And whats the best storage media if the data has to be stored for a long time? This week we cover Hard Disk Drives (HDDs) and Solid-State-Drives (SSDs)

HDD and its development over time

As previously said in our first part of this article series, tape has a one major drawback: It is a linear storage media. This means that it needs always some time to reach the correct place on the tape were the data is located before it can be read out. This is the main reason why the hard disk was invented. Actually it was the same producer who invented the hard disk nearly 62 years ago, who was also developing the tape media during those ancient computer times: IBM. The introduced the IBM 305 RAMAC mainframe computer on September 14th, 1956. This was the first computer that was shipped with the IBM 350 Disk Storage Unit that was equipted with two internal hard disks. The disks however were quite big: They were each 24 inch in size and each could only store 4,375 MB in total. The mainframe computer was not cheap: Their customers had to pay 3200 US Dollars. Each month! Because IBM did not sell this product but they were for lease. As was the hard disk drive: storing data on this first commercial hard disk drive cost an additional 750 US Dollars a month. The System was followed by the IBM Ramac 1301 six years later in 1962 and its users were now able to store up to 28 MB of data.

But with the invention of smaller devices – the Personal Computer - in the late seventies of the last century, the need for more storage space was increasing. While in the first years of the PC, data was commonly stored on data compact cassettes (they looked like normal audio compact cassettes and hold magnetic tape inside the cassette to store data onto ), this changed in favor of small hard disks.

The first HDD could hold around 5 MB of data and was introduced by a company which was later renamed into Seagate Technology in 1980. That was the beginning of the HDD success story. The Seagate ST-506 was the first hard disk drive in the 5,25 inch format which is still used today. From 1980 until 1992 the storage space such a hdd could hold rose from 5MB to 2 GB. Nowadays HDDs filled with a hellium gas can store up to 14 TB of data.

Simply put a hard disk drive is an electromechanical storage device that uses mangentic storage to store and retrieve digital information from one or more rotating platters coated with magnetic material. The ferromagnetic material on thin film on top of the platter changes its direction when it electricity is supplied. It then represents a „1“ or a „0“ and can be read out. All these zeros and ones represent a binary data bit and an encoding scheme determines how the user data is represented by the magnetic transistions.

The main benefits of hard disk drives are that they have become quite cheap and hold a vast amount of data compared to their price. With the introduction of RAID, hard disk drives are used even more since it is now possible to not only randomly store data on a bunch of independent disks, but also have one or more spare disks to recover data because of a HDD failure.

Most producers give an estimated life span of a magnetical HDD of 5 years. However this depends on its usage: Most HDDs can easily store data for 10 years, but users should frequently check if their data is still available and error-free. If the HDD is not frequently used and only stores data, there is still a good chance that the data is still there after more than 10 years. However it is possible that data can be lost due to the fact that the magnetism vanishes after many, many years.

SSDs: What is the difference?

SSDs however are totally different: They consist only of electronical chips put on a circuit board. Usually a typical SSD holds one controller chip and several storage chips. The controller chip controls how the data is stored on the storage chip. Therefore this chip stores the needed „logic“ how this is done. Another difference is that the storage chips only last for certain read and write cycles. Simply put the more you write the faster the SSD will be broken. This is the reason why most of the manufacturers have inplemented technologies inside the SSD (controller) chips to distribute the data write load onto all storage chips equally. That is due to the fact that if „just“ one oft he many storage chips reaches his End-of-life you can throw away the complete SSD.

The same is true to the much cheaper USB sticks that also hold these somewhat cheap NAND Flash chips. They, too, only last a certain time, while their read/write life span is usually much less compared to SSD and depending on the quality and price.

They too can store data up to 10 years, while as said earlier this only when not used or only read out. You can check your devices life span quite easily by using special tools which you can download for free at many manufacturers websites.

Even though both disk types can reach a 10 year life span they differ a lot when it comes to data recovery. While hard disk are around for many years, so are the data recovery tools and processes professional data recovery service providers use. Other the years they have gathered so much knowledge on how to recover data from a certain hdd brand, that if it is possible to recover data they will do so. With SSDs however, the situation is different. The main problem here is not that the technology is challenging, but that every manufacturer can buy huge amounts of different controller and storage chips, put them on a platter and then program and insert a „distribution code“ on the controller chip.

This said, it is „easier“ to recover data from a wellknown SSD brand than from some asian me-too-product, where sometimes working and specialized recovery tools might not work a 100%. Additionally it is not wise to use SSDs for storage anyhow, because of the fact that they are more expensive compared to HDDs, harder to recover and much faster. Using SSDs for storage and not solely for computing is like having a Ferrari and driving only up to 20 mph on the street.

Next week we discuss optical storage media as well as the cloud as possible storage for longterm archiving purposes. ´Til then…