How is computer server maintenance and data center maintenance changing? What innovation is being incorporated to lower work re
Delivered July 26, 2019. Contributor: Michelle A.
In order to better help the military and other companies to prepare, document changes in the computer server maintenance and data center maintenance space.
With advances in technology, there have been major shifts to storing data in the cloud. As a result, equipment and funding has been taken from data centers, which can make it challenging for managers to continue to maintain the remaining systems that have not been shifted to the cloud.
Because of the growth of big data and the need to store it efficiently, there is a move to decentralize the power of computers and place hardware closer to the locations that need the data. This is being referred to as "the edge," and it is happening because it can be inefficient to stream so much data to the cloud for processing. This will result in "micro-data centers, branch locations, and smaller hubs to process data." which will mean that companies will have to shift where their staff is located to ensure that all the data hubs are supported.
Disc storage will be overtaken by SSD which means that the staff in charge of storage will have to be up to speed on the newer technologies.
Hybrid cloud systems are continuing to grow and Gartner predicts that by 2020, 90% of organizations will have adopted "hybrid infrastructure management capabilities."
The design of data centers will be changing to be more efficient and functional. This includes AFM, DCIM, liquid cooling, and microgrids.
Machine learning has tremendous potential to help with server and data center maintenance by changing from scheduled maintenance to predictive maintenance. By using data to predict when there will be a problem, maintenance can happen only when necessary to avoid potential problems.
Another area that machine learning can help data centers is with energy efficiency. By having components such as outside air temperature, the amount of power being used by the system and air pressure, cooling the data center can happen automatically which reduces the need for human intervention and increases energy efficiency, since needed adjustments can happen in real time.
Google Case Study
In 2014, Google started using machine learning to improve energy efficiency in ts data centers. The model was designed to make recommendations of what changes should occur and human then implemented the changes.
The company has now switched to a fully automated system where the change happen automatically in real time. This system is saving the company about 30% per year on energy costs for the data center cooling system.
The process has taken time because the more data collected, the more efficient the system can become. The company first had to gain confidence in the recommendation the system was making. Once that happened, it was easier to automate more.
Only the project owner can select the next research path.