According to Gartner analyst Dave Cappuccio, 80% of enterprises will have shut down their traditional data centers by 2025, compared to just 10% today. The figures are fitting considering the host of problems faced by traditional data centers. The solution to all these problems lies right in front of us- Incorporating Intelligence in traditional data centers.
To support this claim, Gartner also predicts that by 2020, more than 30 percent of data centers that fail to implement AI and Machine Learning will cease to be operationally and economically viable.
Across the globe, Data science and AI are influencing the design and development of the modern data centers. With the surge in the amount of data everyday, traditional data centers will eventually get slow and result in an inefficient output. Utilizing AI in ingenious ways, data center operators can drive efficiencies up and costs down. A fitting example of this is the tier-two automated control system implemented at Google to cool its data centers autonomously. The system makes all the cooling-plant tweaks on its own, continuously, in real-time- thus saving up to 30% of the plant’s energy annually.
Source: DataCenter Knowledge
AI has enabled data center operators to add more workloads on the same physical silicon architecture. They can aggregate and analyze data quickly and generate productive outputs, which is specifically beneficial to companies that deal with immense amounts of data like hospitals, genomic systems, airports, and media companies.
How is AI facilitating data centers
Let’s look at some of the ways that Intelligent data centers will serve as a solution to issues faced by traditionally operated data centers.
#1 Energy Efficiency
The Delta Airlines data center outage in 2016, that was attributed to electrical-system failure over a three day period, cost the airlines around $150 million, grounding about 2,000 flights. This situation could have been easily averted had the data centers used Machine Learning for their workings.
As data centers get ever bigger, more complex and increasingly connected to the cloud, artificial intelligence is becoming an essential tool for keeping things from overheating and saving power at the same time.
According to the Energy Department’s U.S. Data Center Energy Usage Report, the power usage of data centers in the United States has grown at about a 4 percent rate annually since 2010 and is expected to hit 73 billion kilowatt-hours by 2020, more than 1.8 percent of the country’s total electricity use. Data centers also contribute about 2 percent of the world’s greenhouse gas emissions,
AI techniques can do a lot to make processes more efficient, more secure and less expensive. One of the keys to better efficiency is keeping things cool, a necessity in any area of computing. Google and DeepMind (Alphabet Inc.’s AI division)use of AI to directly control its data center has reduced energy use for cooling by about 30 percent.
#2 Server optimization
Data centers have to maintain physical servers and storage equipment. AI-based predictive analysis can help data centers distribute workloads across the many servers in the firm. Data center loads can become more predictable and more easily manageable. Latest load balancing tools with built-in AI capabilities are able to learn from past data and run load distribution more efficiently.
Companies will be able to better track server performance, disk utilization, and network congestions. Optimizing server storage systems, finding possible fault points in the system, improve processing times and reducing risk factors will become faster. These will in turn facilitate maximum possible server optimization.
#3 Failure prediction / troubleshooting
Unplanned downtime in a datacenter can lead to money loss. Datacenter operators need to quickly identify the root case of the failure, so they can prioritize troubleshooting and get the datacenter up and running before any data loss or business impact take place. Self managing datacenters make use of AI based deep learning (DL) applications to predict failures ahead of time.
Using ML based recommendation systems, appropriate fixes can be inferred upon the system in time. Take for instance the HPE artificial intelligence predictive engine that identifies and solves trouble in the data center. Signatures are built to identify other users that might be affected. Rules are then developed to instigate a solution, which can be automated.
The AI-machine learning solution, can quickly interject through the entire system and stop others from inheriting the same issue.
#4 Intelligent Monitoring and storing of Data
Incorporating machine learning, AI can take over the mundane job of monitoring huge amounts of data and make IT professionals more efficient in terms of the quality of tasks they handle.
Litbit has developed the first AI-powered, data center operator, Dac. It uses a human-to-machine learning interface that combines existing human employee knowledge with real-time data. Incorporating over 11,000 pieces of innate knowledge, Dac has the potential to hear when a machine is close to failing, feel vibration patterns that are bad for HDD I/O, and spot intruders. Dac is proof of how AI can help monitor networks efficiently.
Along with monitoring of data, it is also necessary to be able to store vast amounts of data securely. AI holds the potential to make more intelligent decisions on – storage optimization or tiering. This will help transform storage management by learning IO patterns and data lifecycles, helping storage solutions etc.
Mixed views on the future of AI in data centers?
Let’s face the truth, the complexity that comes with huge masses of data is often difficult to handle. Humans ar not as scalable as an automated solution to handle data with precision and efficiency. Take Cisco’s M5 Unified Computing or HPE’s InfoSight as examples. They are trying to alleviate the fact that humans are increasingly unable to deal with the complexity of a modern data center. One of the consequences of using automated systems is that there is always a possibility of humans losing their jobs and being replaced by machines at varying degrees depending on the nature of job roles. AI is predicted to open its doors to robots and automated machines that will soon perform repetitive tasks in the datacenters.
On the bright side, organizations could allow employees, freed from repetitive and mundane tasks, to invest their time in more productive, and creative aspects of running a data center.
In addition to new jobs, the capital involved in setting up and maintaining a data center is huge. Now add AI to the Datacenter and you have to invest double or maybe triple the amount of money to keep everything running smoothly.
Managing and storing all of the operational log data for analysis also comes with its own set of issues. The log data that acts as the input to these ML systems becomes a larger data set than the application data itself. Hence firms need a proper plan in place to manage all of this data.
Embracing AI in data centers would mean greater financial benefits from the outset while attracting more customers. It would be interesting to see the turnout of tech companies following Google’s footsteps and implementing AI in their data centers. Tech companies should definitely watch this space to take their data center operations up a notch.