Topology: Services for Business Observability

Building on Broadcom’s innovative domain tools, the AIOps and Observability team’s thinking and use of topology has advanced significantly in recent years.

To illustrate these innovations and their benefits for IT operations, we continue in this blog where the previous blog left off. In this post, we cover services—and the services layer—as an extension to AIOps topology. This allows us to achieve service observability and, in the process, business observability.

Below, we cover observability drivers to clarify problems, goals, and benefits for enterprises. We outline how observability adds a maturity level to traditional monitoring, especially when it adds causal and proactive observability.

We also reflect on problems addressed by business-aware observability through service analytics delivered by AIOps and Observability from Broadcom.

Observability drivers

The pressures to achieve new cost efficiencies for IT operations, deliver outstanding user experiences, and gain deep understanding of complex situations are first-order drivers of observability. In our digital era, as business stakeholders turn increasingly to IT as a strategic partner, business-aware observability is emerging as a critical need.

The role of IT: From alignment to blending

The role of IT and its relationship to the business is evolving. These teams move from alignment, as an arms-length provider of services to the business, to an integral partner that shapes customer experiences, business operations, and core business models. Still, large gaps exist between IT and business—gaps that can be reduced by blending IT and business data to elevate understanding.

By addressing these gaps with service observability, enterprises can increase transparency and accountability, and foster proactive ownership by IT of results that matter to business, such as measuring and optimizing digital experience, improving IT resource utilization, and preempting disruption of value delivery chains (all of which have a digital component).

For this evolution, unified, business-aware observability is required. The challenge: how to blend, in a single pane of glass, IT operations data with business context in a way that is accurate and intuitive for the various teams who need this understanding to effect business delivery and performance?

Closing the gap between IT outputs and business outcomes

The desire to close the gap between IT and business, to align IT outputs and business value, has a long history of modest results.

With modern practices like Agile, SRE, DevOps, cloud, and CI/CD, which allow technology teams to operate more autonomously, closing the gap is becoming more challenging. Enterprises move more quickly, have more data and more organizational silos, and are more distributed—rather than being centrally controlled. Without strong alignment and correlation with business KPIs, teams may be optimizing discrete areas of IT but failing to improve holistically. A simple example: Improved performance of a cluster or database may have no bearing on what matters to business value delivery, although it might help a team achieve their SLA.

How do businesses know which IT or software development projects deserve investment when they cannot connect IT outputs to business outcomes?

AIOps and Observability from Broadcom can help align and correlate IT outputs with business outcomes to help technology teams work and collaborate constructively with line-of-business teams. Instead of a narrow, inward-facing focus on IT execution, with service analytics, teams can begin to prioritize work and measures of performance focused on business outcomes.

Improving IT productivity fast

With the right information presented to them, in a context meaningful to their IT domain, IT teams can make this transition quickly—and happily, since this context-rich information can improve confidence in their decisions. Service topology helps to correlate and contextualize IT operations data.

Importantly, users of all experience and skill levels must be supported with the right insights to take appropriate actions. Relevant information enriched with IT context and business context will help them work efficiently within their domain team. In addition, it will help them to work collaboratively and constructively with other IT domain teams and with line-of-business stakeholders.

Whether on an individual employee-, team-, cross-team level, IT estates (or ICT infrastructures) are too complex and interconnected for teams to fully grasp. The realm of important contexts is simply too vast for individuals or teams to embrace. This is where powerful, innovative technology for services is needed.

The measures of IT success

The measures of IT success are changing. The era of declaring success based on five 9’s uptime for individual IT silos or qualitative judgments is over. The language of success within IT now revolves around quantitative measures connected to business value and risk, such as customer experience, service and value delivery, time-to-value, and risk prevention.

For these measures, complete, confident, and unified business-aware observability is required.

Service observability

Service analytics capabilities combined with service modeling and topology within AIOps and Observability from Broadcom are the cornerstones of service observability.

Together, they link IT resources, monitoring data, and measures of system health, performance, and availability. This is all captured by IT operations to reflect business value delivery. When end user monitoring is included, IT and business teams can get the full end-to-end view needed for service observability.

Services within a service layer

To reflect IT and business services as a topology, we use service components, a novel extension of topology. Adding this logical layer elevates IT observability to the business level and forces alignment and correlation of IT outputs to business outcomes.

• Service: A service is a virtual grouping of IT components that participate in the delivery of an IT or business service as a named service. By modeling the dependencies, relationships, and attributes of these components, the solution generates a service topology. With services and the service topology, the solution can group conceptually related subsets of an ICT infrastructure.
  
  Since enterprises have many services, which typically have cross-service dependencies or may be structured hierarchically, the service topology becomes a dynamic and powerful construct to help IT monitor and manage the matrix of IT resources.
  
  The service topology will reflect shared and common IT services and their delivery into business services that would be at the top of the hierarchy.
• Service layer: A services layer is a topology layer for separation of service components.

Services map business services to your topology; they indirectly map your ICT infrastructure to your business functions. As AIOps and Observability solutions understand topology, services enable AIOps to elevate causation to the services layer. This effectively elevates observability to the business level, making the observability of AIOps from Broadcom business aware.

The services layer illustrates the power of layers and the focused perspectives they offer for collaboration. It forces teams (business, operational, technical, and expert alike) to focus on shared views of topological contexts that are important to IT and the business. These may be business services, IT services, application groupings, clusters, and so on—whatever sub-views are most relevant.

Importantly, these service constructs naturally cross IT silos. Conceptually, service topology acts as a “service context bus” that connects relevant teams across IT and carries IT operations data that each team needs for triage and remediation.

Business observability

Service analytics enables business service observability.

• Service observability: With service topology, IT operations can collaborate with business teams to model digital services that they can then actively monitor and manage for better alignment of IT outputs with business outcomes.
• Business service observability: Service observability applied to business services.

AIOps and Observability from Broadcom adds the ability to define SLIs and SLOs for services to enforce business tolerance thresholds, whether they stem from compliance, resilience, or best practice demands.

• Service level indicators (SLIs) are derived metrics that describe chosen aspects of the health or risk for a service. SLIs are based on aggregating metrics for components of the service to indicate availability, error, or latency.
• Service level objectives (SLOs) define tolerance levels of SLIs to form an SLO percentage value that is used to gauge SLI compliance with regards to defined error budgets.

Business-aware operations through operational guidance

Services help provide a holistic overview of a system's health that is business aware and that also includes IT environment dynamics via topology.

When teams understand how IT operations data correlates with business outcomes, such as digital experience, they can correctly prioritize work, without missing opportunities to better manage the health of the environment, address risks, and learn from monitoring insights.

Typically, each IT domain team monitors their domain in isolation from the others. For example, application teams focus on applications. While legacy monitoring can't provide further insights, the correlation with business data improves MTTR and system analysis. With business-aware operations across IT teams, enterprises can improve how they triage incidents, prioritize remediation, and allocate resources. Communication and accountability also improve.

Summary

With services and by defining SLIs, SLOs, and utilizing error budgets, enterprises can elevate observability to an organizational level that extends to business operations. Services enrich raw monitoring data from IT operations, so they can be associated with business outcomes, such as digital experience.

With the cost pressures and competitive forces of the digital era, organizing around digital services is becoming a strategic initiative for many enterprises. Service observability with AIOps and Observability from Broadcom can be a catalyst for this services model to help support digital transformations at a massive scale.

Supported by service topology, each team can better manage IT systems and infrastructures, with immediate understanding of the service context and associated business outcomes.

• Identify the most significant problems and solve them faster with service analytics
• Contribute constructively to end-to-end service health and performance and while focusing on business impact
• Monitor and manage IT resources based on service priority
• Establish SLIs, SLOs and error budgets that align to business needs for service delivery
• Create IT services which directly contribute to business services and correlate infrastructure and application KPIs with business outcomes.

Other blogs in this series

1. Topology for Confident Observability and Digital Resilience
2. Topology for Incident Causation and Machine Learning within AIOps
3. Capturing a Complete Topology for AIOps