Eric Silva

3 AI-based Cures for IT Over$pending

Blog Post created by Eric Silva Employee on Apr 29, 2019

Trying to relax but feeling the $pinch$ from the Finance department? IT overspending is a problem shared by many businesses, especially ones growing or in highly competitive markets. Luckily today there are advancements in data center management technologies that can tackle some of the biggest IT challenges businesses face when growing rapidly or under cost containment mandates. While data center management maybe the most risk averse bunch in any organization, there are technologies based on AI, advanced analytics and machine learning that can do the work that humans just can’t possibly do. Once the reward outweighs the risk, even the most risk averse IT and business executives are willing to jump in. How about saving over 30% on your cooling costs? Maybe now I have your attention.

 

1 – Data Center Cooling Optimization

While AI can be found all over the data center it is often hidden behind the management user interface for a device or built-in to the device itself, providing the core value. If we could take AI and rise it above the siloed technology level to support the entire heterogenous data center, that is the level of impact that can offer the rewards risk averse and cost conscious IT executives are looking for. One such advanced AI-enhanced technology is cooling optimization. Cooling the data center is a costly endeavor and it's not a simple one either. The data center landscape is constantly changing; servers, applications, and business units coming and going.n large data centers there are many variables that affect the temperature at any given spot in the room. The temperature in a data center can also be impacted by user behavior. For example quarter end financial reporting can increase need for power and cooling across multiple applications and tens of servers. It is quite common for data center management to waste hours adjusting cooling systems, chasing data center hotspots in order to optimize the cooling and reduce the thermal risk those hotspots create. Optimizing cooling of large, complex data center is really next to impossible for a human to do.Thanks to advanced analytics, rack sensors, and machine learning, a cooling optimization engine can analyze all the spots in the data center at once. It can then estimate what the impact would be of adjusting various fans and chillers to perfectly optimize the data center temperature. This lowers thermal risk, extends the life of expensive IT equipment and dramatically lowers IT utility costs creating a greener data center:

sdc 4.png

2 - SAN Optimization Across Heterogeneous Storage

Data centers today have SAN optimization tools made available to them from their SAN networking hardware vendors. They do a good job managing the network on that group of devices, but what is needed is to be able to optimize multiple SANs and manage them across the entire data center. The ability to manage storage at a higher level then that of a single vendor’s SAN optimization tool would enable a broader view and the ability to apply predictive analytics on multiple SAN performance. The predictive analytics can then suggest or strongly recommend actions to enhance performance and meet demanding SLA’s across all applications and heterogeneous SANs:



3 – Predictive IT Budgeting

 

From refrigerators that you can inventory while at the supermarket to floor cleaning robots, consumers are getting used to smart devices and robotics. The complexity of data centers and the thousands relying on them, makes levering artificial intelligence, IoT and automation a bit more challenging than washing your kitchen floor. There are so many different vendor technologies to manage and IT technologies across all the business units, analytics are very dynamic, hard to define and harder to analyze. It’s no wonder IT organizations across the globe are so risk averse when it comes to IT automation. Managing data centers is like managing your own NASA space program, no one else is doing exactly what you are doing. A data center is made of many moving parts, none quite the same and there isn’t a guide book that comes along with it. There are many vendors involved with many inter-dependencies. These are not simple challenges and it’s difficult for an IT organization to tackle them all at once or go it alone. Smart Data Center from Hitachi Vantara has pre-engineered common IT automated tasks that can be quickly tailored to your unique data center and ITSM (IT service management) tools. Along with intelligent automation, Smart Data Center provides valuable budget, performance and capacity estimates that enable accurate views into the future from the device, application to business unit requirements. Being able to predict at the business unit and device level enable much more accurate budget forecasting and the ability to prevent unforeseen service outages:


So there’s 3 quick ways to leverage AI for IT overspending while staying risk averse and maintaining SLAs. There is no reason to wait or feel like you are implementing AI on the bleeding edge. While it seems it’s going to be a while before the average data center can go on “auto-pilot”, Hitachi Vantara is helping our customers accelerate their use of AI by leveraging predictive analytics and IoT to lower costs and risks while optimizing IT performance. That road and vision is a platform to build on, available today: Smart Data Center from Hitachi Vantara - AI for the data center.

Outcomes