Machines can now read the news and have distinct advantages over humans in this regard. For one, they don't get hot under the collar paging through the paper. Plus, they can discern patterns, learn what to look for and react to information - all in microseconds.
The idea that news drives the ups and downs of the stock market is long established. Bloomberg has terminals at each of their traders' stations to deliver the latest news driving market sentiment, both broadly and on specific stocks. What has changed in the past few years is the speed of news and the volume of data – humans simply can’t keep up. Events-driven trading is just one example of an industry being upended and put back together by AI. News reports travel globally at internet speeds throughout the day with any one of them able to affect market outcomes for a specific vertical or across industries.
While humans can’t keep up - neither can the traditional model of the static data center. The computational power of GPUs is changing how we architect alongside ultra-fast NVMe storage and evolving CPU and networking technologies. A core problem remains: workloads rarely look alike. Keeping every combination of hardware available to find the right ratio of resources is rarely cost effective or consistent. Adding to this complexity is the array of specialized hardware being released over the next 24 months to outperform generic compute resources for specific algorithms.
The only solution is to move away from static infrastructure fixed at the point of purchase to a dynamic architecture that can adjust on demand. This is where Liqid enters the picture. Composable Infrastructure enables the placement of physical assets where and when they are needed. For Edge computing, where the bulk of information processing will take place, this means being able to check out the hardware from resource pools as needed and then release those resources when idle or a better option is available. This allows AI-driven infrastructure to evolve with the data.
Estimates on the global impact of AI range widely — from a possible $46 billion by 2020 to a potential $15.7 trillion by 2030. Over half of the data generated will be handled in Edge facilities outside traditional data centers by 2020. A recent IDC analysis notes that spending growth on AI is expected to be in the double digits across all industries with regulated markets such as banking and healthcare among early growth drivers. It's clear that no sector or business is immune from the impact of AI.
What is certain, as Dylan and others have famously sang - “the times they are a-changin'.” We are going to stop pushing data to where the infrastructure is and instead push infrastructure to where it needs to be - on the Edge, in real time, where it matters most.