Real-world uses

How open source is used in e-commerce, PaaS, and HPC machine learning environments

on Jul 14, 16 • by Kara Howson • with No Comments

A recap of our recent webinar on real-world uses for open source...

Home » Open Source » How open source is used in e-commerce, PaaS, and HPC machine learning environments

“People think they can just dive in head first but it’s actually a paradigm shift.” Justin Reock, Rogue Wave Software.

From Apache and OpenSSL to Docker and Kubernetes, true understanding of your open source environment comes with experience. In this week’s webinar, the third in our open source webinar series, we covered plenty of real, industry-spanning experience as seen by our “example” organizations: UniTrack (mobile PaaS), PupBand (e-commerce website), and HelpingHumans (HPC machine learning).

Using these three real-world examples, we walked through identifying the infrastructure needs, the technology stack selection process, and the final architected solution for each organization. Be sure to watch, re-watch, or share this on-demand version of the webinar to see how these companies successfully used open source software.

Interesting facts

Throughout the webinar, we asked attendees questions to learn more about how they use open source in their organizations.

Poll: What percentage of your mission critical software is open source?

What percentage of your mission critical software is open source?

Our expert Justin was surprised to see that the 51-75 percent adoption wasn’t higher given that we are seeing such broader adoption rates and Gartner predicted that 99 percent of Global 2000 companies would use open source this year. However, it’s good to remember that open source use isn’t always known in an organization, therefore the numbers on questions like this can actually be skewed.

Poll: Which of the packages/technologies that we just discussed are you using?

Which of the packages/technologies that we just discussed are you using?

It was great to see such a huge number of OpenSSL and PHP users.  However, Justin recommends those that haven’t looked into Camel to do so. It’s a really amazing integration platform and a straight forward way to do both event driven and static integration. The component library is also very impressive.

Poll: In your opinion, how different are containers/microservices vs. service-oriented architectures? 

In your opinion, how different are containers/microservices vs. service-oriented architectures?

Our host Greg said that he would fall into the “don’t know” and he’s was not alone. As Justin explained:

“A lot of people look at this concept of microservices/containers as a buzz word and really the answer here is there is a very big difference. What is different is that we have new technologies and innovation/advancement in container and automation, and the supporting technologies that make it possible for us to deploy modular code in a way that is way more efficient than it was. The concepts are not new in terms of modular development, it’s just the technology is now there to allow us to do it in a way that wasn’t possible before.”

Answering your questions

The session was very interactive, and we received some great questions which we promised to recap.

Can you provide more explanation around Kubernetes and is it like Docker Swarm?

[Justin] It is similar to Docker Swarm in that it is solving the same problem however it’s development is further along. They solve similar problems – sophisticated management of your container environments, discovery of new container-based services that are coming up, orchestration and elastic scale, automated deployment, and manipulation of containerized environments. Swarm is still considered to be experimental, not production ready, but release is pretty close. Kubernetes is just more mature and used inside Google, and by Netflix and as part of their elastic, cloud scale architecture.

The last example used TensorFlow, can it be made to work with Spark?

[Justin] Spark can be used for data analytics, but TensorFlow lends itself to machine learning analytics as opposed to just general data analysis. TensorFlow and Spark can be used together, where Spark can provide maintenance on large data sets, and the analyzed data can be provided to TensorFlow to perform learning on optimized data sets.

Which of the stacks mentioned would be the most stable in terms of uptime?

By providing redundancy through multiple elastic scaling containers, the Docker solutions will likely provide the highest level of system availability. Extending that, management of the Docker containers through Kubernetes is the current most reliable solution, and so theoretically the HelpingHumans stack would provide the highest level of reliability and availability.

What’s next?

Watch, re-watch, or share this on-demand version of the webinar – Open source applied: Real-world uses

Catch up on the webinars you missed:

– Part one: How enterprises learned to stop worrying and love open source

– Part two: When is free not free: The true costs of open source

Register and plan to attend our fourth webinar on July 27, when we dive deep into specific OSS packages to examine the top issues in the enterprise with two of our most qualified OSS architects, Bill Crowell and Vince Cox.

If you have more questions, leave us a comment. We’d love to hear your thoughts on open source.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *

Scroll to top