Organizations can derive extraordinary value by collecting new data smartly and efficiently, in the right place at the right price, and analyzing that data to win business gold.
Is there real gold in those hills? Since I live in Colorado, I think of that question whenever I go west into the Rocky Mountains. Driving through abandoned mines and tourist spots where families can search for gold I can’t help but think there’s still untraceable gold to be discovered.
But what is the cost and risk of getting that gold? It would be hard enough to locate it, and even harder to extract and take advantage of. At some point, gold mines from early eras realized that “the juice wasn’t worth squeezing anymore” and closed shop. Nevertheless, my eager curiosity makes me wonder how much gold has not yet been discovered.
Today, mining takes place in mountains of newly generated data, extracting nuggets of valuable and profitable information. But at what cost? Poor performance and higher costs? There has to be value in all that data, right?
Data gravity can take its toll on an organization. With proper planning, organizations achieve extraordinary value by collecting new data smartly and efficiently, in the right place at the right price, and analyzing that data to win the gold. To do this, we need to collect, refine and reserve.
Collection: The big picture of the global new data creation world, according to Statistaindicates that by 2025 we will be generating 180 trillion gigabytes of data. Today, most of that data is NOT stored (~2% or 1.28 ZB), but that percentage is expected to grow to over 25% by 2025. Combining data creation growth and data storage growth, that’s an increase of 3,086% of the amount of data stored from 2020 to 2025. Like a snowball rolling down a mountain and getting faster and bigger, the impact of data gravity – unresolved – is vastly increasing.
A common process to minimize your data footprint is to implement data reduction technologies such as deduplication and compression.
Refinement: In algorithm-based decision-making processes, new data comes in and is then analyzed against rules set by training with previous incoming data. As more data is collected, new patterns are identified, generating more accurate and personalized results. The algorithms are updated and the value proposition increases. Where should this refinement synthesis take place? At the location where it was made? Or somewhere else so that it doesn’t burden the environment needed to deliver the highest throughput?
To reserve: Lifecycle management policies are critical to the data gravity discussion. We collected, we refined, but now what? Will this data be of any use, or have we extracted all of its usefulness? The answers will most likely differ depending on how the data is used. Want to compare satellite images over time? Saving image data can be useful on the go. Have you taken the average fast data recording speed? Can you delete the data once you have received answers? Probably. Perhaps it never needed to be saved, especially if it was possible to measure the average ingestion rate of data in real time. The point is that different data types have different values and should be managed in the best possible way, with the least amount of pain possible, especially in terms of cost and latency.
Extracting gold from ore requires a process called gold cyanidation. It is a dangerous endeavor due to the toxic nature of cyanide, requiring expensive measures to process and dispose of these harmful chemicals. The methods used and their costs are measured by the income from the gold. Do they slow things down, affect business, or are they too expensive, decreasing value to the mining company?
As data gravity increases, so does the negative impact on the business, requiring significant changes that can reduce business-constraining latency and data costs.
The sheer number of abandoned mines in my beautiful home state tells me the answer to those seeking those veins of gold. Most likely, the juice was no longer worth squeezing.
For more information, visit Dell.com/Edge