A Brief History of Data: Part 1 - Infrastructure
From hieroglyphics to the cloud to punch cards, find out in this article how data infrastructure has evolved over the centuries and millennia.
Among the Egyptians
The number of hieroglyphs reached up to 5000 different signs. Findings suggest that the Egyptians used a classification system for hieroglyphs. The evidence seems to be too tenuous to fully understand this early system for classifying thousands of hieroglyphs.
Let's jump back in time to the end of the 19th century. The founder of IBM, Herman Hollerith invented the punch card in 1890. They allowed the census in the United States in 1890. These machines worked with a tabulation table and did not require a computer. They were complex to use.
The storage capacity of punched cards was then 960 bits. They were successful until the invention of hard disks.
Then IBM came back in 1956 with the first hard disk. The first one weighed almost a ton and was 1.8 meters high. It was called RAMAC 305 (Random Access Memory of Accounting and Control). By grouping together 50 magnetic disks with 50,000 sectors, it could store 3.75 Mb of data. Ridiculous today, it was an unprecedented push for the time. We could now record data.
Then miniaturization started and thus data storage in quantity could really begin. Today's hard drives are up to 16TB, a far cry from the capacities of 1956.
The servers then made it possible to centralize the storage and processing of data and to distribute them to different users within the same entity. Companies, administrations and research laboratories were thus able to pool resources to manage the different needs of users and drastically increase computing power and storage requirements according to usage.
We are therefore moving from the individual or group of individuals who manage their computer to a centralization of storage and operations, which reduces costs and provides more flexibility in terms of needs.
In 1999, Salesforce was the first private company to offer a cloud storage service followed in 2002 by Amazon with AWS (amazon web services). The cloud makes it possible to do the same thing as servers. But it is no longer on the scale of an entity but on the scale of the whole world or often of a continent - for performance reasons, geographical proximity is preferred. Cloud users will benefit from on-demand storage and computing capacities.
It's interesting because if a user needs to temporarily increase his power by 100 times, he doesn't need to buy 100 times more servers or machines that he already owns and that he won't need afterwards. The cloud will allow him to rent this power just for the time he needs it. And we don't have to manage physical machine resources. A service provider takes care of that for us.
Nevertheless, today the cloud raises questions about data governance. How can we be sure that data shared on the Cloud will not be shared with third parties? Is the data well protected? Potentially better than on a server, because we have also centralized the security competence. Delegating the management of the machines to a third party implies that you lose some control over the processing of the data.
Today, most applications are stored in datacenters, in the cloud.