IT departments are central to all digital transformation strategies, but they are in need of help, due to the existing complexity of infrastructure and application footprints.
63% of organizations reported that time spent on correlating data from different sources is the key challenge for usability of IT performance monitoring tools1. This reflects more in business downtime and outage costs. Companies end up spending 72% of their budget just "keeping the lights on" instead of focusing on innovation and spending time increasing the value of their business.
The average organization contains at least five IT monitoring tools, all doing specific but different tasks, and each producing its own siloed data. This huge amount of data is superficially unrelated and disorganized. There is still no holistic view of the IT environment available in a context that businesses can understand.
To deal with the problem, we need the help of AIOps (Artificial Intelligence Operations).
AIOps platforms use big data, machine learning and other advanced analytics technologies to enhance IT operations (monitoring, automation and Service Desk) with proactive, customized and dynamic awareness. Instead of replacing all of your existing tools, which can potentially take years, AIOps takes over the pain of correlating disparate data to automate repetitive processes. This will give your organization free time to work on revenue generating IT projects and ensure that IT service delivery is optimized.
AIOps is defined as: "Software platforms that combine big data and machine learning functionality to enhance and partially replace a broad range of IT Operations processes and tasks, including IT monitoring, event correlation and analysis, IT service management, and automation."
The emphasis of AIOps platforms is in the ability to collect all types of data from various IT management systems, in a scalable and rapid way. It then applies business context and smart automation on that data to enable your IT teams to be smarter, more responsive and proactive. The key benefits of AIOps are:
Prevent downtime and improve service delivery
Move from siloed teams to IT business partners
Automate tedious manual tasks
Develop your IT maturity
Achieving a successful digital transformation via an AIOps strategy can be a journey for most organizations. With an end goal of becoming an IT business partner, it is necessary to consider these three focus areas:
Savision advises a six step approach to implementing a successful AIOps solution. The end goal should be a holistic platform that integrates with all existing IT Management tools. To achieve this, it is important to start small by adding value with a single integration, provide improved visibility for one tool at a time, improve service monitoring and off er automation. Once that value is proven, the next phase of the deployment should be to add further data-sources.
AIOps platforms should be fully extensible to allow scalability in depth and breadth across your IT organization. The main requirements for a platform should include:
Savision uses Elastic Stack to provide this big data platform. Built on open-source technology from Apache Lucene, this is a well-established enterprise class technology that provides the speed and scalability to manage huge data-volumes and remain flexible enough to handle data-sources from any tools that generate log fi les.
The last piece needed to provide a great search facility for IT organizations is the ability to look at related data in a common way. Users will not want to interpret operational alerts in seven different ways because it came from seven different tools that each have a different format. There needs to be a consistent way of interacting with common data types such as: alerts, incidents, computers, websites, databases, etc.
Savision has over 10 years of experience in the IT Operations world, working with tools such as Microsoft SCOM, Solarwinds, and Nagios. By applying business context to alerts (coming from components and grouping them into business services), you can assess business impact and focus on issues that are affecting end-users.
Savision developed a universal data schema that can store any data point collected dynamically from an IT organization, but also normalize those data points so that they can be easily managed. For instance, Savision is able to present all computers and alerts in a common way even if they are managed from completely different IT management systems.
A key feature of Savision's AIOps platform is to provide value to users from IT administrators, Helpdesk, or IT managers. Data visualization is critical in order to provide context on alert noise issues that come from a single monitoring tool.
Savision is able to quickly build custom dashboards to visualize your IT environment and share with stakeholders on any device. Technical dashboards are automatically created based upon standard business applications (e.g. Microsoft Exchange, Active Directory, CRM, Order Entry), together with high-level dashboards for any persona.
Savision Dashboarding features include:
IT Operations run many business and custom applications over virtualized infrastructure. These are the tangible pieces that traditional monitoring tools are built to actively maintain. However, IT teams are ultimately responsible for delivering IT services such as Email, Telephony, CRM etc. These services, are relying on a set of components that all need to function in order for the service to be up and running. With a traditional mindset, you would need to monitor all the components individually to make sure they have a healthy state. But where's the relationship between these components?
This is where service-oriented monitoring comes into the picture. Instead of watching a single component, the complete service should be monitored with all its underlying components. This can be done within AIOps, where we can visualize these relationships. For your web shop to function properly, you need the following components to work:
When speaking of service-oriented monitoring, there are three layers that need to be defined:
Once you have defined these layers, you can start monitoring the complete service.
Other important features include:
Now that we have created situational awareness from dashboards and service focused monitoring, the next step is to automate processes to save time for IT departments.
Savision is able to filter out alert noise and focus on real incidents to become proactive. The next step is to automate incident workflows to speed up MTTR (Mean-Time-To-Resolution).
Once a single integration has been deployed, the model can now be reproduced across all other IT management systems. The holistic view or single pane of glass can be rolled out throughout your multiple tools and IT teams. Integrate all of your IT monitoring, IT service management, public, and private cloud data into a single pane of glass. Integration packs for all major IT systems include:
The final part of an AIOps deployment can deliver great value to your business over time. With all data-sources ingested and normalized into a central database, it is now possible to make use of machine learning capabilities.
"Time Series Anomaly Detection" is a capability used by unsupervised machine learning. This approach builds a steady baseline of your IT operations overtime. When anomalies appear that are outside of normal operations, an incident is created automatically. Automation rules can then kick in if the anomaly matches a similar previous event by creating immediate remediation of the issue.
AIOps is a transformative technology for the enterprise today. The process is a journey and should be approached in a modular method. Focus should be put on building out value on every single IT Management tool. Over time an AIOps platform can spread throughout the organization, eventually with machine learning and automation taking over many of the manual tasks that administrators perform today. A successful implementation can assist a Digital Transformation strategy and increase the IT maturity of any enterprise to become an IT business partner.