Businesses are constantly generating more and more information without adequate budgets to accommodate the influx....
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
IDC expects data volumes to double every two years and reach 40 zettabytes in 2020. While data is increasing at double digit speeds, information technology budgets are increasing at a snail's pace: IDC projects that IT spending will increase 2% and generate $3.8 trillion in 2016. As a result, virtualization technicians often find themselves inundated with information. In response, vendors like VMware are turning to artificial intelligence to ease that workload.
For example, VMware CEO Pat Gelsinger noted in his keynote speech at VMworld 2015 that artificial intelligence is pushing the IT world to new heights. For VMworld 2016, artificial intelligence (AI) is a focus point in its "Content Catalog" for end-user computing. So what is artificial intelligence and how can it improve system management? Dave Schubmehl, a research director of Cognitive Systems and Content Analytics at IDC, said AI software is often thought of as what is seen in movies like 2001: A Space Odyssey and The Matrix.
The term has garnered distorted media attention that often muddies its discussion. The underlying AI features are found in a growing number of products. Some, like the speech recognition available in mobile devices, are familiar. Others are less familiar, like machine learning, where algorithms based on large data sets learn and predict future system needs and proactively meet them.
Drowning in data
Those AI tools can potentially offer help to overwhelmed virtualization center technicians. As corporations have embraced virtualization, applications run on fewer devices, generate more information and increase workloads. Automation has always been a high priority for virtualization managers and has become even more important as systems have become denser and more complex. With AI, software offloads repetitive, tedious tasks and allows IT pros to focus on more important chores, like system and network design.
AI software is emerging as a possible way to streamline a series of virtualization duties.
"AI is a series of rules and heuristics that solve specific tasks," Schubmehl said. Companies generate many alerts -- system performance, network intrusions and workload completion -- and need simpler ways to interpret them. Data centers are becoming too complicated for traditional engineering and human intuition to be manage these processes quickly and efficiently.
In response, vendors are developing algorithms that collect performance information, outline possible problems and either recommend or take action to streamline the remediation processes. Data center virtualization management tool vendors like VMware, Aragon, BMC Software, CA Technologies, IBM, Oracle and Microsoft are developing such products.
Being highly energetic
Energy usage is another area where interest in AI software is high. The Natural Resources Defense Council, an environmental action organization, expects that American businesses will pay $13 billion for data center electricity in 2020. Vendors like Brocade, Intel, Opengate Data Systems, Schneider Electric, SynapSense and Vigilent have focused on improving cooling systems so they automatically adjust to temperature, pressure and workload changes in the data center.
These products rely heavily on machine learning to improve data center efficiency by moving beyond what engineers can analyze. For instance, Facebook developed software that addresses the tendency for onboard server fans to spar with row-level cooling systems as the temperature rises. Its load balancer redistributes workloads across servers and shifts compute activity away from hot spots forming inside racks.
AI financial tools are also on the docket. Romonet, a United Kingdom-based maker of data center management tools, developed a software program -- previously known as Prognose -- that uses machine learning to build predictive data center cost models. The product outlines data center total cost of ownership. The tool takes design and engineering documents and builds simulations that outline how well -- or poorly -- a facility will support new workloads.
AI quickly moves into the mainstream
Such features are being integrated into a growing number of applications. According to IDC, AI software will be incorporated into about half of all applications developed by 2018. IDC also predicts that, by 2020, AI savings -- reduced people costs and increased workflow efficiencies -- are expected to total about $60 billion for U.S. enterprises.
As the technology takes hold, new challenges arise. AI software is complex and difficult to develop. While it relies on automation, somewhere a human being writes the code, so there will still be a margin of error and inefficiencies might arise.
As a result, many businesses do not feel comfortable handing system operations over completely to software applications. The products do not automate everything, but instead mix machine and techie interactions. Typically, virtualization managers are given recommended actions, and they decide whether to accept or reject them. Consequently, there is a learning curve, as techies become comfortable with the tools' capabilities and shortcomings.
Historically, AI has been the work of science fiction movies. Lately, the technology has been working its way into virtualized systems. Moving forward, AI promises to help business offload tedious tasks and enhance productivity, so techies do not drown in data.
Is artificial intelligence a threat to humanity?
Artificial intelligence makes its mark on the enterprise
The danger of AI in the intelligence arms race
The future of AI depends on human assistance