I’ve spend Day 2, attending lectures and talking with participants and sponsors. It was my last day at KubeCon + CloudNativeCon 2019 because I had to return to Warsaw for a meeting of VMware User Group Poland that was scheduled for Thursday.
Once again, I’ve started the day with a keynote. What interested me the most was a session run by Bryan Liles. Brian said, that Kubernetes is not a destination but a journey that takes us to it. What an excellent thought! And one that is right on the money. A platform to build other platforms. I share a similar opinion. Technology should be treated as a way of addressing a goal. It is a tool, and put in the capable hands, it gives amazing possibilities to all users. Here are some key takeaways from Day 2:
- Kubernetes is a Journey Not a Destination,
- you need to plan for failure by having redundancy and the ability to do roll-backs,
- migrate large complex infra gradually,
- have a culture of learning, not blame.
Some people I’ve managed to talk to, brought up the subject of SRE vs DevOps. It was interesting to hear, that SRE are being compared to special task forces, because of their high skills and expertise. Ken Owens from Mastercard told me about their journey towards Cloud Native. Like everyone else, Mastercard worked with few competence silos: storage, networking, sysops, virtualization, etc. After a while, they’ve merged all departments into one, and reinforce it with additional programmers. When I asked if they regarded them all as SRE, he said no. The unified department implements a Cloud Native Developers doctrine and SRE work beside it. In other words, programmers know how to create development environment by themselves without the outside help, whereas SRE handle the maintenance or perform special tasks. Mastercard wants every member of CND department to be self-sufficient. If need be, programmers get an assistance from SME (Subject Matter Expert) or from Architects from their own teams. I’ve asked Ken if he sees a potential danger in the growing number of abstracts that “make things easier”, but in fact lead to the creation of a very complicated stacks. He said that he does, and his main concern is the inadequate number of specialists proficient with specific technologies. That is why he supports the idea of an interdisciplinary team that can spread the knowledge acquired internally by its individual members. Personally, I think it is a very interesting and important project, albeit extremely hard to carry out. In principle, it forces two worlds to collide – worlds of generalists and specialists – and thus confronts different views and ways of solving problems. Many companies follow similar path, hoping for favorable end results. I’m really curious what will come out of this “devopsed” process that is, after all, being implemented on a really gigantic scale.
Next, I’ve attended a session held by Stephan Fudus and David Meder-Marouelli from 1&1 Mail & Media Development & Technology GmbH called „Benefits of a Service Mesh When Integrating Kubernetes with Legacy Services”, concerning Istio service mesh. It was a neatly done introduction to Istio, explaining what it is, what it addresses, and what challenges it poses. Although the idea behind the project is right, one could see that the presented 1.1 version is still in the early stages of development and its implementation causes many problems. Let us all hope it will provide the code with a more democratic access to infrastructure. But is it, at this point, a production grade software? I don’t think so. It is, however, worth observing because of its abilities, its rapid growth, and the fact, that it addresses community’s needs. It’s a good moment to mention Jeff Beker’s words. During a discussion panel titled „Leveraging Cloud Native Technology to Transform Your Enterprise”, Jeff was asked about the right path leading towards Cloud Native. His answer really stuck in my mind. He said, „you won’t get that level of innovation from single vendor or provider like from community”. Unfortunately, I didn’t have an opportunity to talk with Jeff about his thoughts on the matter, because I don’t think that most of the firms are ready for that extent of innovations. They would have to familiarize with them first. Then, find them a proper use. And finally, adopt them. It is the scale of innovation, its vastness and rapid growth, that creates obstacles. Many companies have problems settling with an AWS offer, not to mention taking a step beyond. But I do hope to delve into this interesting idea in the foreseeable future.
Day 2 of KubeCon + CloudNativeCon 2019 ended with a party held on Montjuïc hill inside Poble Espanyol. Beautiful scenery encouraged casual talks. Pity, it was my last day at this more than interesting event. I’ve met a ton of wonderful people from different groups and communities. I’ve asked them the same questions. What was the purpose of your visit? Why did you come here? Some came to learn about K8S ecosystem, others to see new innovations. I’ve talked with some lecturers and met old friends, who wanted to discuss very specific subjects, like storage in cloud native environment, SDN, or SD-WAN. I’ll try to attend this event again because it opens one’s eyes to a new approach to IT, and innovations fueled by the community. Today, its Warsaw and PLVMUG meeting I am co-hosting. I just can’t wait. In the meantime, see you somewhere around the world!