Why machine learning is important

 　　Beginning with Lee Sedol vs AlphaGo in the Go world in 2016, artificial intelligence (including machine learning) instantly broke into our vision. Cell phone photography, artificial intelligence can help you automatically identify the scene and select the best shooting parameters, saving you post work; fraud risk control, artificial intelligence can automatically identify risky transactions, thus preventing the loss of user assets, as well as self-driving cars, intelligent voice customer service, etc.

　　Today all kinds of products or software claim to have built-in artificial intelligence. Although they are used in various areas of production and life, though the core idea is the same, to help free people from simple labor, whether in the office or otherwise, and free up more time to deal with other problems.

　　This philosophy is also being used in the storage space.

　　

　　From installation and deployment to operations and maintenance management, enterprise storage has become increasingly intelligent and automated in recent years, with a gradual reduction in manual involvement. Administrators only need a simple operation, the storage array can be automated to complete the initial configuration; management functions have also become more intelligent, through intelligent analysis and optimization features, storage arrays can automatically find the system bottlenecks, for the administrator to make recommendations.

　　In addition, with the introduction of artificial intelligence technology, enterprise storage is gradually evolving from a machine that executes commands to an autonomous system capable of self-learning.

　　However, not all storage companies have this condition, because for any artificial intelligence, the three elements of data, computing power and algorithms are necessary for AI to achieve remarkable success. The AI has to have a lot of data to support it, and it needs to cover all possible scenarios so that it can get a model that performs well and can be more intelligent. And after having data, it also needs to be trained, constantly.

　　

　　Therefore, to introduce artificial intelligence into the storage field, the test is a company's strong R & D strength and deep industry experience, after all, the storage is stored in the most valuable assets of enterprises - data.

　　As the industry's premier end-to-end IT solutions provider, Dell Technologies, with decades of deep IT infrastructure cultivation and industry-wide customer accumulation, was the first in the industry to launch PowerMax, a high-end storage with built-in machine learning engine, in 2018.

　　PowerMax leverages predictive analytics and pattern recognition to automatically place data on the right media based on I/O profiles (without incurring additional overhead), resulting in greater performance improvements.

　　

　　As an example, a typical 200TB PowerMax array can analyze and predict 40 million data sets in real time, driving 6 billion decisions per day, including.

　　1. where to place the data (flash or SCM)

　　2. what data should be compressed or deduplicated

　　3. which QoS service levels require more performance

　　In June, Dell Technologies launched another new midrange storage product, PowerStore, where machine learning is used in a broader context. powerStore is a set of federated storage clusters up to 4 Appliances (each Appliance is a dual control node storage).

　　

　　*Dell EaseUS PowerStore is powered by Intel® Xeon® Scalable processors, which optimize workloads with high reliability, high compute power, high stability and efficient agility, helping PowerStore not only meet established workloads with ease, but also prepare for digital change.

　　So which Appliance is the best choice for a business system to put data volumes on? In the past, this has been a huge headache for storage administrators. Here's a real-life example.

　　An enterprise-class user has five sets of storage deployed in their IT environment. Each time a new business comes online, the storage administrator needs to decide which storage to deploy the business data to based on the performance and capacity requirements of the business, so the storage administrator needs to have a full and continuous grasp of the operational status of each storage and make accurate judgments.

　　However, business needs are constantly changing, and when the business runs for a period of time, it is likely that the original storage is no longer suitable for that business, and then a new storage device has to be selected for it.

　　On the other hand, data migration has also become a new problem. The risk of extensive manual operations during data migration and the impact of downtime is a minefield that no administrator wants to touch.

　　But in PowerStore, everything is different.

　　With PowerStore's machine learning engine, it automatically analyzes which appliance is the best choice for deployment in a cluster based on the storage requirements of the business, and keeps an eye on the growth trend of data volumes in real time while the business is running, and automatically notifies administrators when a migration is needed and gives them the best migration recommendations.

　　When the administrator confirms the operation, PowerStore will automatically migrate the data among the Appliances in the cluster, and the switching process requires no downtime. Through such a mechanism, data can always run in the most suitable storage environment to ensure the quality of data services.

　　

　　Today, the sheer volume of data generated by humans and computers is far beyond the capacity of humans to absorb, interpret, and make complex decisions based on it. Artificial intelligence forms the basis of all computer learning and represents the future of all complex decision making. To master AI is to master the future, and a storage system with built-in AI can help you stay one step ahead on the path to "mastery".