Portworx this week announced it has made possible to migrate both applications data and pod configurations between Kubernetes via a PX-Motion tool that has been added to version 2.0 of the company’s PX-Enterprise storage software.
In addition, the latest version of PX-Enterprise includes PX-Central, a tool for monitoring application data on a Kubernetes cluster.
Portworx CTO Goutham Rao says PX-Motion will enable IT organizations to run multiple instances of Kubernetes clusters as well as more easily back up those clusters by moving data to, for example, another instance of Kubernetes. Portworx integrates with Kubernetes services running on AWS, Microsoft Azure, Google Cloud, IBM Cloud, Pivotal, Red Hat OpenShift, Mesosphere DC/OS, Heptio and Rancher.
A survey of 519 IT professionals published this week by Portworx finds 82 percent are running containers, up from 55 percent a year ago. A full 60 percent report that more than 40 percent of their applications are now running in containers, the survey finds. One-third of survey respondents report running containers in more than one cloud. Just less than a quarter (23 percent) are running in two clouds and 13 percent are running in three clouds.
The survey also finds the top two reasons organizations are embracing containers are to increase developer efficiency (32 percent) and a desire to more easily take advantage of multiple clouds (30 percent). Overall, Microsoft Azure (57 percent) was the most popular cloud for running containers, followed by Amazon Web Services (AWS) at 42 percent and Google Cloud platform at 39 percent. Well more than a third (36 percent) are running container in all three major public clouds. The most popular combination of two clouds is AWS + Azure (27 percent) , followed by Azure + Google (24 percent), and then AWS + Google (13 percent). More than half (54 percent) are also using two or more container schedulers.
Rao says that as Kubernetes gets deployed both in the cloud and in on-premises IT environments, the challenges associated with managing Kubernetes become that much greater. In fact, survey respondents rank security, data management along with IT operations spanning multiple clouds are the top three barriers to enterprise container adoption.
The survey also finds 36 percent of respondents are now spending more than $500,000 on container license and usage fees per year. A total of 12 percent say they are investing more than $1 million.
Rao says data and storage management associated with containerized applications is a challenge on multiple levels. Containerized applications tend to scale up and down in a way that requires access to storage systems that are designed to scale out because the application is no longer confined to a single set of virtual machines. The number of containers running on any given server can also put a lot of pressure on I/O as hundreds of containers seek to access a common pool of storage, he adds.
Storage is becoming a bigger issue as IT organizations begin to deploy stateful applications on Kubernetes. The challenge now is figuring out how best to optimize the performance of those applications as they access both local and external storage resources.