How teams are breaking down data silos to improve software delivery

Over the past decade, the software development landscape has shifted dramatically. Long development cycles resulting in big releases are now a thing of the past. Instead, the focus is on delivering value to the customer quickly and efficiently through small releases delivered at a high frequency.

This shift is not without its challenges. Increased release velocity means an increased threat to security and reliability. To combat this, teams must collaborate fluidly across the organization. As part of Sumo Logic Illuminate 2021, Carl Caum, Senior Product Manager at Sumo Logic, and Ian Buchanan, Principal Partner Engineer at Atlassian discussed these challenges and how data and visibility help organizations overcome them.

So, what is needed? “Harder, Better, Faster, and Stronger” software delivery

Software development processes have evolved significantly as businesses scramble to keep up with the competition by delivering value to their end-users more efficiently. But simply delivering software quickly isn't enough. As Carl Caum says, these expectations can be represented by the phrase, “harder, better, faster, stronger.” Let’s break that down.

Harder

The software that organizations deliver must be secure, without slowing down the rate at which software is delivered. To achieve this, security needs to be “built into every single thing that you do—from code commit to delivery and operations,” says Caum. "Security is not an afterthought, but a critical part of the development process."

Better

The software also needs to be better, meaning it needs to function as designed and be highly reliable. Customers today have numerous digital options in virtually every industry. If a particular product is unreliable, end-users will quickly find an alternative.

Faster

The end goal for development teams today is to deliver value to customers more quickly than ever. This requires an emphasis on automation and a concerted effort to continuously optimize the development process.

Stronger

In today’s development climate, collaboration across teams is a must. Effective collaboration serves to eliminate bottlenecks in the development process and facilitates the release of high-quality products with greater frequency.

Data to the rescue

So we know what we want from the development process, but how do we get there? One of the key tenets of DevOps is to measure everything to attain the visibility necessary to identify areas that need improvement. This principle holds when addressing the productivity challenges that lead to delays in delivery.

Atlassian's Buchanan uses the term “inactivity” to describe this slowness, identifying it as the “silent killer” of efficient software delivery. This inactivity needs to be measured.

If it takes two days to move between steps—your automated testing won’t identify the dead time in the process—but your team needs to be aware of it.

For instance, an individual tool that acts within a delivery pipeline will likely provide measurements and timeframes for what it does. Most teams know how long it takes to run their automated tests. Taking an ideal example, let’s suppose it takes five minutes to run automated tests. However, if it takes two days to move beyond this step, your automated testing tool won’t identify the dead time in the process—but this is information that a team needs to know to understand bottlenecks and improve processes. In other words, broader metrics that encompass the entirety of the development cycle are necessary for ensuring efficient collaboration and fast delivery.

Key metrics to measure—the advantage of elite performers

The State of DevOps 2019 Report, DORA (DevOps Research and Assessment) discusses how to properly measure the effectiveness of a software delivery pipeline. More directly, they examine how “elite performers” stack up against “lower performers” in these categories.

• Deployment frequency - Elite performers deploy code 208 times more frequently than lower performers.

• Lead time from commit to deploy - Elite performers are 106 times faster from commit to deploy than lower performers.

• Incident recovery - Elite performers recover from incidents 2,604 times faster than lower performers.

• Deploying breaking changes - Elite performers deploy breaking changes 7 times less often than lower performers.

As the stats indicate, being an elite performer gives you a significant competitive advantage. On the flip side, if you aren’t an elite performer and your competitor is, this represents a risk to your business.

It's easy to see why these metrics are crucial to track. But since the typical delivery pipeline is composed of a series of fragmented and siloed tools, how can this be done? The answer lies in leveraging the proper platform that collects and correlates data from disparate tooling to provide teams with useful insights to evaluate and improve delivery performance.

Using Software Development Optimization to drive efficient software delivery

To bring together and make sense of the data from complex and fragmented delivery pipelines, organizations need progressive and comprehensive monitoring solutions. Sumo Logic Software Development Optimization provides the insights necessary to advance the development process.

Software Development Optimization

Software Development Optimization (SDO) is a unique solution for compiling critical delivery performance measurements to drive improvement in software delivery.

To help break down long-standing data silos, SDO comes with out-of-the-box integrations for popular pipeline tools. Upon integration, data is collected from these tools and then aggregated and analyzed to provide DevOps teams with context as to how their delivery pipeline is performing as a whole. SDO normalizes and correlates data from the various tools that make up the organization's pipeline, allowing for visualizations that represent a real-time view of the DORA metrics discussed above.

This aggregation of metrics across tools gives teams the necessary data to evaluate delivery performance more completely. Furthermore, SDO drills into these critical delivery performance metrics to isolate the culprits that negatively impact delivery efficiency more easily. This allows DevOps teams to focus on the portions of their process that are suffering the most.

For the many organizations that leverage Jira for collaboration, SDO is available as a free service directly within the JIRA cloud from the Atlassian Marketplace. To check it out, you can request early access to SDO in the JIRA cloud.

Conclusion

The expectations for modern software development teams continue to rise as organizations strive to deliver high-quality software in shorter time frames than ever before. The right tools and processes are needed to meet this Herculean task, helping teams to break down silos and build better software faster.

To learn more about Software Development Optimization, please reach out and request a 30-minute demonstration.