Edge computing could shift deeply ingrained business practices if it produces enough additional value.
For example, take the standard practice sending personal data to the cloud, which most organizations do to extract personalized insights produced by computation-heavy machine learning.
Privacy is the issue, including data breaches and compliance with government data privacy laws.
Edge-level analysis protects privacy to a greater extent, while still extracting insight for personalization. Consumer personal data remains more secure without sacrificing behavior analysis. Personalization still happens but associated legal and reputational risks, latency and costs are reduced.
The amount of new value versus cost is part of the thinking. Edge processing might be preferred for some apps and business processes, even if the cost is higher than the cost of hyperscale data center processing.
One always can get into a debate about the value versus the cost. The cost of processing at the edge is almost certainly going to be higher than doing so at a hyperscale data center, on a per-byte basis.
But the total cost--including bandwidth cost--can produce different outcomes, as analyst David Floyer first argued in 2017.
In some ways, the value versus cost debate will be similar to the debate over storage media.
It is often argued that data centers will switch to flash storage, simply because--ignoring performance advantages--flash storage might be less costly than the hard disk drive alternative by about 2027.
One can often get a debate over the benefits and costs of storage using flash versus disk drives in data centers, even if 90 percent of the time, observers agree that flash higher performance and higher cost makes sense for some applications while HDD, with lower performance and lower cost, still makes sense for less critical apps.
In other words, flash might always make sense for tier 0 for apps such as financial transactions and e-commerce apps. Flash might make sense for some tier 1 business processing, data mining, and data analysis tasks as well, though not necessarily.
Most agree that HDD is fine for tier 2 for functions such as email, file and print, data archives, and backups.
But not everyone believes flash actually will become cheaper than disk drive storage, at least not for some time.
Hard disk drive technology advances will continue, some argue. The issues might include how much improvement can happen, and whether flash might continue to have a higher cost per unit of storage than HDD, and if so, for how long.
But you would be hard pressed to find even proponents of HDD arguing flash will “never” become cheaper than HDD for data center storage.
On the following chart, the orange line in this log-scale x-axis chart is the hard disk drive (HDD) $/TB street cost over time, plotted against the left-hand x-axis.
The blue line represents the street $/TB SSD cost over the same time period, and the two lines are projected to cross in 2026. The green dotted line stands for the ratio between the SSD and HDD costs, plotted against the values on the right vertical axis.
source: Blocks and Files
No comments:
Post a Comment