Building with Containers Stateless or Stateful… that is the Question

New year, new challenges…

Clearly last year was the year of the Docker – it finally made it into the Enterprise world and that’s great news. Following a few steps behind, Kubernetes continued to strengthen its position for container orchestration for typical stateless, micro-services based Docker Apps; adopted by most lead vendors including key competitors.

Stateless vs Stateful Building with Containers Infographic

Stateless vs Stateful Applications - New year, new challenges… Clearly last year was the year of the Docker - it finally made it into the Enterprise world and that’s great news. Following a few steps behind, Kubernetes continued to strengthen its position for container orchestration for typical stateless, micro-services based Docker Apps; adopted by most lead vendors including key competitors. More and more folks are now wondering if this modern approach can also accommodate their pre-existing apps, whether as old as 90s-2k Dot.Net or as recent as the web-scale distributed Java apps in VMs - either way, containerizing is one thing, adhering to the stateful aspects is another. Clearly, with modern web, edge, IoT and alike large scale - Docker and Kubernetes were best optimized for stateless, not for stateful. Retrofitting stateless into a stateful architecture is not a walk in the park. In an attempt to stop-gap, Docker has introduced Docker Storage Plug-in which could be handy for very simple micro-service apps. HA and scale remain an issue even in this base use case. Challenging the model with clustering and distributed data proves to be difficult even with straight forward collection of containers some mistakenly call a cluster. The next couple years will be interesting from an adoption point of view - bringing in the zillions lines of code and apps into the modern container world is clearly a big challenge many break their teeth on. Luckily, Robin is set to rescue, focused in/around the most complex cases featuring Big-Data and NoSQL, as well as “legacy” RDBMSs such as Oracle which for now only support LXc.

More and more folks are now wondering if this modern approach can also accommodate their pre-existing apps, whether as old as 90s-2k Dot.Net or as recent as the web-scale distributed Java apps in VMs – either way, containerizing is one thing, adhering to the stateful aspects is another.

Clearly, with modern web, edge, IoT and alike large scale stateless use cases – containers could be best optimized. Retrofitting stateless containers architecture into a stateful model is not a walk in the park.

In an attempt to stop-gap, Docker has introduced Docker Storage Plug-in which could be handy for very simple micro-service apps.  HA and scale remain an issue even in this base use case.  Challenging the model with clustering and distributed data proves to be difficult even with straight forward collection of containers some mistakenly call a cluster.

The next couple years will be interesting from an adoption point of view in the stateless vs stateful world – bringing in the zillions lines of code and apps into the modern container world is clearly a big challenge many break their teeth on.

Luckily, Robin is set to rescue, focused in/around the most complex cases featuring Big-Data and NoSQL, as well as “legacy” RDBMSs such as Oracle which for now only support LXc.

mm

Author Razi Sharir, Vice President Products

More posts by Razi Sharir, Vice President Products
Site Info | Disclaimer | Privacy Site Network | Add to favorites © 2009-2018 PHP Beautifier - All rights reserved Related terms: programming, coding, php, scripts, dev, programmers, code, Zend, Pear, javascript