Telemetry data holds the key to flawless, secure, and performant digital experiences

Organizations need to build complete customer-centric environments that deliver superb, secure, personalized digital experiences every time, or risk losing out in the race for competitive advantage. Prioritizing both internal- and external-facing applications and ensuring they are running optimally is the engine behind every successful modern business.

The complexity of cloud native and distributed systems has risen in lockstep with the expectations of customers and end users. This rachets up the pressure on the teams responsible for applications. They need to aggregate petabytes of incoming data from applications, services, infrastructure, and the internet and connect it to business outcomes.

This telemetry data — called MELT or metrics, events, logs, and traces — contains the information needed to keep digital experiences running at peak performance. Understanding, remediating, and fixing any current or potential breakdown of the digital experience depends on this collective data to isolate the root cause.

Given our dependence on performant, real-time applications, even a minor disruption can be costly. A recent global survey by IDC reveals the cost of a single hour’s downtime averages a quarter of a million dollars — so it’s vital that teams can find, triage, and resolve issues proactively or as quickly as possible.

The answers lie in telemetry, but there are two hurdles to clear

The first is sorting through vast volumes of siloed telemetry in a workable timeframe. While solutions on the market can identify anomalies, or issues out of baseline, that doesn’t necessarily mean they are a meaningful tool for cross-domain resolution. In fact, only 17% of IDC’s survey respondents said current monitoring and visibility options are meeting their needs, though they are running multiple solutions.

The second is that some data may not even be captured by some monitoring solutions because they see only parts of the technology stack. Today’s applications and workloads are so distributed that solutions lacking visibility into the full stack — application to infrastructure and security, up to the cloud and out to the internet where the user is connected — are missing some vital telemetry altogether.

Effective observability requires a clear line of sight to every possible touchpoint that could impact the business and affect the way its applications and associated dependencies perform, and how they are used. Getting it right involves receiving and interpreting a massive stream of incoming telemetry from networks, applications and cloud services, security devices, and more, used to gain insights as a basis for action.

Cisco occupies a commanding position with access to billions upon billions of data points

Surfacing 630 billion observability metrics daily and absorbing 400 billion security events every 24 hours, Cisco has long been sourcing telemetry data from elements that are deeply embedded in networks, such as routers, switches, access points and firewalls, all of which hold a wealth of intelligence. Further performance insights, uptime records and even logs are sourced from hyperscalers, application security solutions, the internet, and business applications.

This wide range of telemetry sources is even more critical because the distributed reality of today’s workforce means that end-to-end connectivity, application performance and end-user experience are closely correlated. In fact, rapid problem resolution is only possible if available MELT signals represent connectivity, performance, and security, as well as dependencies, quality of code, end-user journey, and more.

To assess this telemetry, artificial intelligence (AI) and machine learning (ML) are essential for predictive data models that can reliably point the way to performance-impacting issues, using multiple integration points to collect different pieces of data, analyze behavior and root causes, and match patterns to predict incidents and outcomes.

Cisco plays a leading role in the OpenTelemetry movement, and in making systems observable

As one of the leading contributors to the OpenTelemetry project, Cisco is committed to ensuring that different types of data can be captured and collected from traditional and cloud native applications and services as well as from the associated infrastructure, without dependence on any tool or vendor.

While OpenTelemetry involves metrics, events/logs and traces, all four types of telemetry data are essential. Uniquely, Cisco Full-Stack Observability has leveraged the power of traces to surface issues and insights throughout the full stack rather than within a single domain. Critically, these insights are connected to business context to provide actionable recommendations.

For instance, the c-suite can visualize the business impact of a poor mobile application end-user experience while their site reliability engineers (SREs) see the automated action required to address the cause.

By tapping into billions of points of telemetry data across multiple sources, Cisco is leading the way in making systems observable so teams can deliver quality digital experiences that help them achieve their business objectives.


Additional Resources

Learn more about Cisco Full-Stack Observability

Read further on future-proofing observability with OpenTelemetry


Carlos Pereira

Cisco Fellow

Chief Architect, SI&A