How HPE InfoSight AI proactively spots, solves infrastructure issues
Find out how the cloud service uses AI, machine learning and predictive analytics to collect and analyze data from millions of sensors to address problems before they happen.
Data center infrastructure is more complex than ever, making it susceptible to service disruptions and extended downtime. In such an environment, IT can have a difficult time identifying the root cause of a problem, let alone preventing the problem from occurring in the first place.
Hewlett Packard Enterprise (HPE) attempts to address these challenges with HPE InfoSight, a cloud-based service for proactively managing data center systems. It collects data from millions of sensors and uses HPE InfoSight AI, machine learning and predictive analytics to continuously analyze the information to help customers address potential infrastructure issues.
The HPE InfoSight AI difference
HPE offers a number of storage platforms that target application workloads. For example, HPE 3PAR StoreServ midrange all-flash array scales from a few terabytes to more than 80 petabytes in a four-system federation. But flash drives alone aren't enough to ensure that an array is properly managed and optimized. IT teams must continuously monitor and assess their storage systems to keep them running at peak performance and maximize their availability.
To help with this process, HPE offers the InfoSight service to its customers at no charge so they can more effectively manage their data infrastructures. HPE InfoSight AI technologies are able to help predict and prevent issues across the infrastructure stack. In addition, InfoSight can help optimize performance and maximize resource use.
HPE's 3PAR StoreServ storage systems
InfoSight pulls data from HPE storage and server systems across the globe, including data from virtual environments based on VMware technologies. After analyzing the data, InfoSight can predict potential problems and automatically generate resolutions, which are passed onto the customer.
HPE has been collecting system data for more than a decade, while continually expanding data collection capabilities. InfoSight uses machine learning to correlate data across infrastructure stacks and then match known problems to customers with similar infrastructure configurations.
InfoSight uses predictive analytics to help prevent and automatically resolve infrastructure-related issues, addressing a range of concerns, such as storage performance, capacity, server security and system availability issues. It continually monitors every system throughout the worldwide installed base, using application and resource modeling to constantly learn and improve.
InfoSight management
InfoSight adheres to an AI-driven management model that simplifies resource allocation and administration. The platform provides a centralized portal that enables customers to quickly understand how they can optimize resources, improve performance and plan for future changes. They can also view infrastructure status and health, with insights that assess both past data and also look to the future.
In this way, customers have a single source of truth that spans their entire infrastructure stack. For example, InfoSight might identify performance issues between the infrastructure's virtual machines and its storage systems.
InfoSight collects data from millions of sensors, continuously analyzes the information and uses the results to help customers address potential infrastructure issues.
InfoSight's approach to preemptive management can lead to quicker resolutions, more reliable storage systems and better performing applications. It can also prevent complex problems before they occur and minimize the impact on IT. According to HPE, as much as 86% of potential problems can be automatically predicted and resolved, reducing the number of support calls and time spent investigating and fixing issues.
If an issue does occur, InfoSight can address it faster because the platform has already collected extensive information about the customer's environment, making it possible to get right to the problem's root cause. If the problem can't be resolved automatically, HPE support engineers proactively investigate the issue and contact the customer with a fix, even if the problem is unrelated to storage. The customer doesn't have to recreate the problem or dig up log files or even initiate the call to HPE, resulting in less time and fewer resources needed to manage storage.
The InfoSight process
The InfoSight platform is built on a highly scalable cloud architecture that includes the components necessary to collect, manage, analyze and distribute data. The entire service process is based on four steps: data collection, data processing, recommendation and data delivery.
Every second, InfoSight collects millions of sensor measurements from HPE storage and server products around the world. At the same time, it collects system logs, alerts, heartbeat data and configuration information. However, InfoSight can't collect data until customers have enabled data collection with the InfoSight service. Once this process has been initiated, customers can control which data sets are captured, and they can opt out of InfoSight at any time.
As data is collected, InfoSight passes it through a massive processing pipeline that cleans, transforms and organizes the information and identifies critical relationships and associations. InfoSight also uses a portion of the data to train the data science models, which provide the foundation for predicting issues in customer infrastructures. This process is an ongoing operation in which InfoSight continuously learns from the globally installed base, leading to more precise and accurate predictions.
InfoSight also uses a recommendation engine that builds off the predictions to make actionable recommendations for preventing and addressing customer issues and optimizing their infrastructures. The recommendation engine uses HPE InfoSight AI and machine learning to go beyond simple troubleshooting to identify and prevent more complex issues. For example, the engine might discover ways to improve performance based on I/O workload patterns, taking into account variables that have the highest impact.
InfoSight also makes it possible for the recommendations to be applied automatically. When this approach isn't appropriate, the recommendations can be delivered through support case automation. In either situation, InfoSight can provide IT with methods for improving the infrastructure. IT administrators can also connect to the InfoSight portal directly to view information about their systems, such as configuration data, wellness status checks, data analytics and other useful details.
The InfoSight platform
HPE offers InfoSight as a free service, but only to HPE customers. It isn't available as a stand-alone product. However, for HPE customers, the service could prove a valuable asset, especially since it's already included as part of the product support package.
The main consideration for customers is whether they're comfortable with HPE collecting telemetry data from their systems. If security and compliance are primary considerations, an organization might not be able to fully participate. But for other customers, HPE InfoSight with AI, machine learning and predictive analytics is worth checking out.
