1661879476105 Kubernetes

Utilizing Kubernetes to boost cloud performance

June 16, 2021
Enterprises can create containers where they can run any virtual process or application. 
In recent years, Kubernetes has taken the world of big data by storm. According to original research conducted by my company, Pepperdata, more than 65% of big data workloads will be running in Kubernetes by the end of this year.

Why? Because Kubernetes is the leader in open-source container orchestration. With Kubernetes at the helm, enterprises can create containers where they can run any virtual process or application. These can range from small microservices and software processes to massive applications. Kubernetes is so effective and dynamic at utilizing resources that enterprises can easily break down containers and quickly assign resources to other applications.

This is why Kubernetes is essential for cloud performance, especially for big-data applications. IT teams all need to be doing everything they can to optimize their Kubernetes deployments as best they can.

The world before Kubernetes

Before Kubernetes, services tended to be heavyweight and resource intensive. Activating multiple applications simultaneously would result in performance issues.

To address this, Google decided to create microservices—tiny services to make applications more manageable and scalable. However, because of their size, microservices tended to proliferate quickly. Considering the massive scale of Google's infrastructure, they needed something to manage all these microservices, particularly from a scaling and debugging perspective. Google wrote a predecessor to Kubernetes called Borg. Then, they used what they learned from Borg and developed Kubernetes.

Google released Kubernetes as open-source software in 2014. Since then, Kubernetes has greatly improved reliability and reduced the time and resources needed for daily operations.

Major factors that drive the switch to Kubernetes

Many  IT experts are dedicated to maximizing their use of Kubernetes. More than 30% of enterprises used Kubernetes in some way in 2020. And more than 21% reported that they are seriously looking into Kubernetes to determine how the technology can impact their business processes and goals.

Why are companies so keen to leverage Kubernetes?

1.     Improved resource utilization

A big-data Kubernetes study has found that nearly 77% of organizations adopted Kubernetes to “improve resource utilization and reduce cloud costs.” Kubernetes’ containerization enables developers, testers and other IT professionals to work on large applications and projects without consuming resources.

2.     Accelerated software/application development

Kubernetes enables DevOps and ITOps teams to collaborate through the sharing of software and dependencies in isolation from each other. This process substantially decreases the volume of workloads required for developing and testing multiple versions of software and applications. Some 45% of organizations now leverage Kubernetes for the “development and testing” of new big-data technologies as well as experimenting on new products.

Because applications are written, tested and deployed inside a container, everything is optimized and sped up. This also simplifies the collaboration between teams involved in the process and quickly resolves any issues among developers, admins and testers. Thanks to this streamlined workflow, teams can build new software, roll out new updates, and launch new app versions at an increased pace.

3.     The appeal of open source

Primarily, an open ecosystem like Kubernetes helps industrial enterprises avoid vendor lock-ins because they can test, update, experiment and improve technologies effectively and thoroughly before deploying them for actual use. Secondly, Kubernetes ensures that new technologies from different sources connect and integrate without any friction.

The challenges of Kubernetes

For all its benefits, Kubernetes is a highly complex technology. Companies struggle to effectively and properly deploy and manage their Kubernetes environments while keeping their expenses manageable. It is not a system that you can run on autopilot.

Migrating to Kubernetes is, in itself, a massive hurdle for organizations. Embracing  Kubernetes requires additional supporting technologies and prerequisite skills, which many enterprises fail to invest in prior to migration.

Without the appropriate technologies and framework, or if the skills are not in place before migration, enterprises often suffer from overspending. Our study on big data and cloud applications revealed that one in three enterprises experience a 20-40% increase of their cloud expenses compared to their initial budget. Compounding the migration problem is that 27% of enterprises use manual monitoring methods to track their big-data Kubernetes environments. Another sizable number utilizes generic application-performance-monitoring tools. 

The fact is Kubernetes is far too challenging for manual monitoring, and even general-purpose tools prove inadequate to effectively oversee its resource utilization and performance. Add up these factors and you have a visibility and monitoring nightmare.

What companies need are products that provide visibility and continuous tuning for big-data analytics.

Overcoming challenges through automated big data infrastructure optimization

For organizations leveraging Kubernetes, optimization is a must. This requires solutions that provide visibility into not just applications but also infrastructure for troubleshooting, debugging and planning. IT teams need to continuously tune their stack for optimal performance. This way, companies can maintain business continuity, ensure that applications and workloads meet SLAs, track resource usage, and  deliver clear accountability.

Remember, a lot can go wrong when making the move to Kubernetes. Outdated applications might need to be rewritten. System bottlenecks can prevent you from getting the most value from your data. Be sure to invest in products that provide visibility and automatic tuning to protect you from these risks. To fully maximize Kubernetes, enterprises need to have granular, on-demand visibility into their Kubernetes environments and its performance.

Ash Munshi is CEO with Pepperdata