BMC Helix ITSM Insights architecture

BMC Helix ITSM Insights is a purpose-built, optimized solution that uses machine learning to group and cluster incident data. As a problem coordinator or a service desk manager, you can use the flexible grouping and clustering options to identify problems proactively and correlate incidents.

BMC Helix ITSM Insights has two layers – the user facing components run on BMC Helix Innovation Suite (Converged Platform) and the backend components are a part of AI Foundation services, where a collection of AI/ML services run to support AI Service Management use cases.

The following diagram describes the BMC Helix ITSM Insights components:

BMC Helix ITSM Insights application

BMC Helix ITSM Insights enables transition from traditional ITSM to AI Service Management by turning data into actionable insights by applying intelligence and automation across all data stores. 

BMC Helix ITSM Insights is an application built on BMC Helix Innovation Studio, which consists of the proactive problem management, real-time incident correlation, and configuration UI. This application is a single instance for every BMC Helix ITSM instance. You can access BMC Helix ITSM Insights from Smart IT, which is available on the Converged Application Server.

From BMC Helix ITSM Insights problem coordinators can visualize incident clusters, conduct additional analysis, and seamlessly transition to the problem investigation process in Smart IT. Service Desk and Major Incident managers can use BMC Helix ITSM Insights to accurately identify incidents related to same situation and formally relate incidents as duplicates to streamline the incident management activity.

A GPU or Azure API key is not required to use the Resolution insights feature in BMC Helix ITSM Insights.

If BMC Helix Service Monitoring (Powered by AIOps) is integrated, the PWA incident screens also link to the probable cause analysis (PCA) for faster resolution of issues. For more information, see Performing causal analysis of impacted services  (BMC Helix Service Monitoring (Powered by AIOps) documentation)

BMC Helix Innovation Suite

BMC Helix Innovation Studio has all Service Management products on a single platform that combines the capabilities of Action Request System and BMC Helix Platform. This consolidation allows you to simplify your deployment architecture significantly and to considerably reduce operational costs. For more information, see BMC Helix Innovation Suite architecture  (Action Request System documentation).

AI/ML Foundation services on BMC Helix Portal

BMC Helix Portal is multi-tenant platform that supports Artificial Intelligence/Machine Learning (AI/ML) services and AI stack. All microservices run as containers over Kubernetes. The AI/ML Foundation services are used to manage algorithms and configurations, and to train and manage machine learning models.

Two training pattern types are supported: online training through Kafka streaming and offline training through Job Manager. Both use AI Configuration service for configuration management of algorithms such as hyperparameter tuning. This layer also includes AI/ML services built for various uses cases such as clustering, anomaly detection, probable cause analysis, and predict major classifiers that have multi-product usage.

• Config service: This service is responsible for providing configuration details for the on-demand execution of clustering algorithm which includes hyperparameters for algorithms and user specific requirements.
• Job Manager: This service is responsible for triggering the on-demand clustering service by spawning the container once the request arrives. It also schedules as well as implements limits and queues of jobs since there needs to be a way to limit number of concurrent jobs per tenant.

• Cluster query service and cluster storage: This service provides management APIs for clusters for tasks such as visualizing clusters in the UI.

• Incident webhooks and incident ingestion service: Incidents are streamed from Remedy ITSM Suite to BMC Helix Portal through this service. The webhook generates an automatic notification when a user creates, updates, or deletes incident records.

• Clustering job: This job runs batch clustering algorithms.  It first fetches incidents from Ticket service and then pre-processes it after which it executes the clustering algorithms.  It stores the results in the cluster store.

• Clustering service: As incidents are streaming in, this service forms clusters from streaming incidents using Natural Language Processing (NLP).  The resulting clusters are updated continuously in the cluster store.

• Ticket service: Responsible for accessing ticket data (incidents) in the context of the BMC Helix ITSM user ID that created the algorithm instance.