It is easy in the data storage and data management world and say that it is constantly changing and new technology is being developed every day. Looking back at the history of data storage and data management I discovered that some of today’s “new” technologies had been around (or conceptually around) for a lot longer than I had thought. What has really changed is how we manage data.
The following graph is a representation of the large data storage events If I created this chart to depict market adoption, it would look much different. I’m not going to get into all of the details for each of these, only enough so you get a feel for the change over the years.
Mechanical Punch Card
Herman Hollerith invented the mechanical punch card in 1890. The company founded by Hollerith would eventually evolve into a small company called IBM. Punch cards were made of thick card stock with the dimensions equal to a bank note (7 3/8 inches wide by 3 ¼ inches high). The first major application of the punch card was for the tallying the 1890 census. Compared to doing the computations by hand, this was a huge advancement in terms of productivity and technological advancement. The 1890 Census only took one year to tabulate while the 1880 Census took eight. In order to use the card you needed a punch, a tabulating machine and a sorting machine; no use of an actual computer was required. The cards and machines evolved throughout the years, but when using punch cards, the equipment cost was high and the process for using the cards was very cumbersome. Writing the software programs and getting the cards punched would take several months. In order to use the system, the software program and initial data had to be loaded into the machine each time you wanted to execute it due to the lack of local RAM to retain all of the programs or any of the output data. It was very time-consuming and inefficient. Although we may think of this technology as long-gone, it was very clear that it was still being used in the 2000 United States presidential election.
Magnetic tape, the cold storage king, was invented in 1928 by Fritz Pfleumer. It was originally used to record audio and then expanded to storing data in 1951 with the invention of the UNISERVO I, the first digital tape recorder. The UNISERVO I was implemented in the UNIVAC I computer. The UNIVAC I had an operator console and up to 10 UNISERVO tape drives. The UNISERVO tape drives were able to provide both read and write functionality at a rate of 12.8 KBps and had a storage capacity up to 184 KB. Tape storage has come a long way since it was first invented. Modern, mainstream, tapes (LTO6) can store up to 2.5TB (6.25TB compressed) although there have been some breakthroughs that have gained up to 185TB on a single tape. If stored properly, tapes have a longevity that is difficult to beat.
Tape’s niche is in long-term storage of data that does not need to be readily accessed (aka. cold storage). The main challenges in managing tapes are storage, retrieval of data and maintenance costs. In a perfect world all of a company’s tapes would be kept in a climate controlled building, perfectly stacked and labeled. They would all have current catalogs listing all of the data on each of them. A data center manager would check them routinely for damage and dispose of end of life tapes promptly when they expire. Anyone who manages tape storage is thinking right now that I AM DREAMING!!! The reality is most people in charge of tape storage do not have the time to accomplish and maintain all of those things. If they have an urgent request for data on tape, without assistance, it could take weeks, if not months to restore depending on the amount of data and the condition the tapes are in.
Hard Disk Drive (HDD)
With IBM ruling the computer market for most of the 1950s through the 1980s, it is no surprise that they were the ones to design and ship the first hard disk drive in 1956. The IBM 350 Disk File was designed to work with the IBM 305 RAMAC mainframe computer. This first hard drive weighed about 1 ton and stood almost 6 feet high. The IBM 350 Disk File was configured with 50 magnetic disks which contained 50,000 sectors and stored 3.75MB of data. The IBM 305 RAMAC mainframe computer combined with the IBM 350 Disk File allowed businesses to record and maintain data in real time. They also provided random access to any record and could simultaneously produce output via print or punched cards. This was a great advancement in technology and made data easier to access than ever before. The challenge with the first round of hard drives up to 1980 was the sheer size of the drive.
In 1980 the 5.25 inch drive arrived on the market and significantly reduced the space needed for the drive. This leads into today’s drive which is even smaller, from 1- to 3.5 inches in size, but much larger in storage capacity (3.5 inch HDDs can hold up to 10TB).
The modern hard drive adds a lot of convenience to data management and data storage. High capacity, nonvolatile storage and lower price per gigabyte than the solid state drive alternative. The complaints with the HDD are power consumption, speed and some may say durability.
The first program, CP, which created virtual machines released to the public in 1968. Before its release in order to time share on a mainframe the resources including memory, disk space and CPU cycles, would have to be manually divided among users. Sharing of resources allowed each user to have its own operating system and share the overall resources on the mainframe. Ultimately, this made it easier to use and more secure. If a user experienced a crash, only their operating system would be affected and not the entire system. This first virtual machine is the basis for today’s hypervisor. VMWare, the current market leader in virtualization, started selling a product called Virtual Workstation in 1999 and launched into the enterprise marketing in 2001 with ESX Server and GSX Server. VMWare has continued to develop complementary products ever since.
Virtualization with the advancements from VMware and Microsoft, revolutionized the way data was stored in a data center. With virtualization, data center managers are able to maximize their storage space and reduce their hardware expenses. Servers are now able to run multiple operations on one server versus the single operation the hardware was designed to implement. What does this mean: Faster server provisioning, increased uptime (the ability to easily move a virtual machine from one server to another) and ease of resource management.
Solid State Drive (SSD)
The first solid state drive, Bulk Core, was introduced to the public in 1976 by Datram. It was sized very different from its modern predecessors measuring 19 inches wide by 15.75 inches tall and had a rack-mount chassis. Bulk Core could hold up to 8 memory boards with 256KB of RAM chips. The advantage back then was the relief from moving parts and you did not need constant power to retain your data. The technology behind the first SSD was the base for the modern SSDs which we know. SSDs as we know them today started shipping around 2001. Due to the lack of moving parts, manufacturers are able to shrink the size of the drive.
The advantages with an SSD drive for managing data is the faster storage and retrieval times. They are great for frequently accessed data and operating systems. Other plausible benefits are reduced power consumption, size, noise and the ability to function in more extreme environments.
Software Defined Storage (SDS)
The concept of Software Defined Storage was born in the 1990s as a mechanism for common interoperability and manageability between hardware manufacturers within a network. In fear that SDS would remove the differentiators in the hardware, manufacturers held off on bringing it to market at this time. When virtualization took off so did the application demands of the hardware behind it. SDS was brought in to work with virtualization to manage hardware requirements and allocate the workload. With the addition of SDS data management has become more efficient. Not only is the managing of the data more efficient, but the use of hardware as well.
Although the idea of Cloud Storage was created somewhere in the 1960s, it did not really come to life until the arrival of Salesforce.com in 1999. Salesforce.com enabled the delivery of enterprise applications via a website, something that had not been done in the past. This delivery of an application over the internet opened the doors for other software companies alike. Amazon followed suit in 2002 with internet-based (cloud-based) services including storage. In 2006, Amazon started renting computers for small business to run applications on and with this the cloud as we know it today was born. Since then a variety of cloud providers have entered the space and recently ISO Standards for the cloud have been implemented.
Data management in the cloud removes the responsibility of hardware management from the company and places it in the hands of the Cloud Storage provider, thus reducing hardware costs for the data center.
The world of data storage has come a long way. From managing paper punch cards on expensive hardware, to not having to manage hardware at all.. As you can see a lot of the “new” technologies of today were actually old concepts or technology reborn. So maybe it is good to look into the past to move forward into the future.