The Cloud Foundry Foundation sponsored this podcast.
A number of services on offer for Cloud Foundry architectures have emerged, offering new opportunities — as well as posing new challenges — for service providers.
Steffen Zuber and Michael Lieser, platform engineers for anynines, were on hand to discuss their challenges offering data services for Cloud Foundry platforms during a podcast hosted earlier this month by Alex Williams, founder and editor-in-chief of The New Stack, during Cloud Foundry Summit Europe 2018 in Basel, Switzerland.
Zuber described how their roles have evolved over the past year as platform engineers.
“Last year, we [adopted] DevOps, as we have enabled users to use Cloud Foundry. We have begun to develop data services customers need to for database applications, messaging queues etcetera,” Zuber said. “We focus on the automation to set up these services. So, you have the features from Cloud foundry also for our data services.”
When offering data services with Cloud Foundry, stateless is the operative word, Zuber said.
“Since your data is stateless, you need to store it somewhere. You need to store information about your customer, for example, which you need data services to do,” Zuber said. “We offer a stateful application that runs in the background for stateless applications running on the platform.”
Lieser also described how his role has evolved since he began developing and managing services for Cloud Foundry. “I previously worked on web design,” Lieser said. “I used to develop apps. At anynines, I now work on the backend side.”
Some of the data services anynines integrate with Cloud Foundry include those on offer from BOSH, which anynines uses to deploy Cloud Foundry. BOSH is just one example. “You don’t know what a customer wants until they tell you,” Zuber said.
“So, we offer some basic services that cover 90 percent of the use cases for the cloud Foundry services.”
When building out their own architectures for use by anynines, Zuber described how his team’s concept of data services fit into its DevOps methodology. Zuber described how his team determines whether its application is highly available on the platform or not, since the data services it is built for the need to be available for customers. Care must be taken to make sure data services continue to work for multiple instances and can offer backups and failovers and self-healing capabilities, he said.
“In our case, we have a base architecture we apply to the new service we are implementing,” Zuber said. “We are responsible, of course, for setting up these services to [manage] them. So, we are developing these services and then managing them.”
In this Edition:
1:45: What is the context you’re thinking about data services in the realm of Cloud Foundry?
5:12: How is the term “state” changing application architectures?
9:48: How is that compared to your own experiences?
12:12: Dependency issues and their issues with data services
14:14: The architectures that anynines has built for its data services
18:04: Going forward, what are some of the factors that you’re considering in the market as you build out your own data services platform?