Welcome to our blog! Welcome to our blog!
Big Data and DevOps: The Case for Bringing Them Together

Big Data and DevOps

If you work with Big Data, you might not think DevOps has much to do with you – and vice versa. But you’d be wrong. Here’s why Big Data and DevOps make sense together.

What is DevOps?

You’ve probably heard of Big Data and data analytics – especially if you’re reading this blog. But if you work in the world of data, you might be only vaguely or not at all familiar with DevOps.

So, here’s a quick definition: DevOps is a philosophy of software development and delivery that emphasizes constant communication across the organization. It reflects an effort to streamline software production by removing the barriers that have traditionally separated developers from IT Ops teams and everyone in between.

Big Data and DevOps

An important concept that is closely related to DevOps is the idea of “continuous delivery” of software. Under the continuous delivery model, code is designed, written, tested and pushed into production environments on a constant basis.

Related: Top 5 Blog Picks about Mainframe DevOps Best Practices for 2017

DevOps makes continuous delivery possible because DevOps facilitates constant collaboration between all the different teams responsible for pushing software down the delivery pipeline – as opposed to traditional modes of software production, where long delays tended to occur whenever code was handed off from one team (such as the developers) to another (such as testers) and no one could work in parallel.

Download Now: Bringing Big Data to Life - What the Experts Say

Big Data and DevOps

You’ll notice that the description of DevOps and continuous delivery didn’t mention data. And it’s true that, by most conventional definitions, DevOps is not closely linked to the field of data analytics.

But maybe it should be. If the goal of DevOps is to make software production and delivery more efficient, then including data specialists within the continuous delivery process can be a big boon for organizations working to embrace DevOps – which, by the way, is now a mainstream practice among even the largest enterprises, according to recent analysis.

Big Data and DevOps

After all, despite the exclusion of data analysts from traditional ways of thinking about DevOps, they have crucial contributions to make at all stages of the software delivery pipeline. By integrating Big Data and DevOps, organizations can achieve the following:

More effective planning of software updates.

Most software interacts with data in some way. When you’re updating or redesigning an app, you want to have the most accurate understanding as possible of the types of data sources your app will be working with. And the sooner your developers have that understanding, the better.

For this reason, being able to collaborate with your data experts before programmers even start writing new code can help them to plan updates in the most effective way from a data perspective.

Lower error rates.

Data handling problems can be a big source of errors when software is being written and tested. And the more complex your application and the data it works with, the higher the chance of errors. Finding those errors early in the software delivery pipeline – or, even better, avoiding them in the first place – saves time and effort. (This is the DevOps principle of “shift-left” testing, which emphasizes testing code changes early, in the “left” part of the development cycle.) Strong collaboration between data experts and the rest of the DevOps team is crucial when it comes to finding and fixing data-related errors in an application.

Better consistency between development and production environments.

The DevOps movement emphasizes the importance of making development environments mimic real-world production environments as much as possible. When you’re writing software that works with Big Data; however, that can be challenging for non-data specialists to achieve. The types and diversity of data in the real world can vary enormously. So can the quality of data that software has to work with.

By being involved across the software delivery process, data experts can help the rest of the team understand the types of data challenges their software will face in production. The result of Big Data and DevOps teams working together will be apps whose real-world behavior matches as closely as possible its behavior in development and testing environments.

More accurate feedback from production.

The final part of the continuous delivery process consists of collecting data from your production environment after your software has been released, then using that data to help understand the strengths and weaknesses of the software so you can plan for the next update. This process depends in part on the work of admins, who help monitor and maintain software in production.

But no one is better qualified to analyze production-related data – which can include things like app health statistics (think CPU time, memory usage and so on), the number and location of users and much more – than data experts. By contributing their data analytics skills to the DevOps feedback process, data experts can help make sure that the organization has the best understanding possible of what’s working and what’s not as part of the DevOps continuous delivery chain.

The list of ways that data experts can and should be involved in continuous delivery processes could go on. The overarching message here is: Big Data and DevOps teams can benefit from work together. If you want to make your software delivery processes as efficient as possible by adopting DevOps – and if you’re a large enterprise today, you probably do – don’t forget to include data analysts within the DevOps workflow you implement. Even though the DevOps movement has not traditionally had much of an association with the world of Big Data and data analytics, it should.

Helping data analysts and everyone else on the DevOps team understand one another is easier, by the way, if you take advantage of data integration and automation tools like the ones from Syncsort. They streamline time-consuming data migration and translation processes, and help ensure better data quality so that your IT staff can focus their energy on what matters most – like deriving value from data – instead of on time-draining, tedious processes.

Christopher Tozzi

Authored by Christopher Tozzi

Christopher Tozzi has written about emerging technologies for a decade. His latest book, For Fun and Profit: A History of the Free and Open Source Software Revolution, is forthcoming with MIT Press in July 2017.
0 comments

Leave a Comment

*