{"id":3487,"date":"2018-11-16T14:46:58","date_gmt":"2018-11-16T14:46:58","guid":{"rendered":"https:\/\/www.appservgrid.com\/paw92\/?p=3487"},"modified":"2018-11-17T15:38:13","modified_gmt":"2018-11-17T15:38:13","slug":"foss-project-spotlight-bluek8s-linux-journal","status":"publish","type":"post","link":"https:\/\/www.appservgrid.com\/paw92\/index.php\/2018\/11\/16\/foss-project-spotlight-bluek8s-linux-journal\/","title":{"rendered":"FOSS Project Spotlight: BlueK8s | Linux Journal"},"content":{"rendered":"

Deploying and managing complex stateful applications on Kubernetes.<\/em><\/p>\n

Kubernetes<\/a> (aka K8s) is now the de facto container orchestration
\nframework. Like other popular open-source technologies, Kubernetes has
\namassed a considerable ecosystem of complementary tools to address
\neverything from storage to security. And although it was first created for
\nrunning stateless applications<\/a>, more and more organizations are
\ninterested in using Kubernetes for stateful applications<\/a>.<\/p>\n

However, while Kubernetes has advanced significantly in many areas during the past couple years, there still are considerable gaps when it comes to
\nrunning complex stateful applications. It remains challenging to deploy
\nand manage distributed stateful applications consisting of a multitude of
\nco-operating services (such as for use cases with large-scale analytics and
\nmachine learning) with Kubernetes.<\/p>\n

I’ve been focused on this space for the past several years as a
\nco-founder of BlueData<\/a>. During that time, I’ve worked with many teams
\nat Global 2000 enterprises in several industries to deploy
\ndistributed stateful services successfully, such as Hadoop, Spark, Kafka, Cassandra, TensorFlow and other analytics, data science, machine learning (ML) and deep learning (DL) tools in containerized environments.<\/p>\n

In that time, I’ve learned what it takes to deploy complex stateful
\napplications like these with containers while ensuring enterprise-grade
\nsecurity, reliability and performance. Together with my colleagues at
\nBlueData, we’ve broken new ground in using Docker containers for big
\ndata analytics, data science and ML\/DL in highly distributed
\nenvironments. We’ve developed new innovations to address
\nrequirements in areas like storage, security, networking, performance and
\nlifecycle management.<\/p>\n

Now we want to bring those innovations to the Open Source community\u2014to ensure that these stateful services are supported in the Kubernetes
\necosystem. BlueData’s engineering team has been busy working with
\nKubernetes, developing prototypes<\/a> with Kubernetes in our labs and
\ncollaborating with multiple enterprise organizations to evaluate the
\nopportunities (and challenges) in using Kubernetes for complex stateful
\napplications.<\/p>\n

To that end, we recently introduced\u00a0<\/a>a new Kubernetes open-source
\ninitiative: BlueK8s. The BlueK8s initiative will be composed of several
\nopen-source projects that each will bring enterprise-level capabilities for
\nstateful applications to Kubernetes.<\/p>\n

Kubernetes Director (or KubeDirector for short) is the first open-source project in this initiative. KubeDirector is a custom controller
\ndesigned to simplify and streamline the packaging, deployment and
\nmanagement of complex distributed stateful applications for big data
\nanalytics and AI\/ML\/DL use cases.<\/p>\n

Of course, other existing open-source projects address
\nvarious requirements for both stateful and stateless applications. The
\nKubernetes Operator<\/a> framework, for instance, manages the lifecycle of a
\nparticular application, providing a useful resource for building and
\ndeploying application-specific Operators. This is achieved through the
\ncreation of a simple finite state machine, commonly known as a
\nreconciliation loop:<\/p>\n