Cloud Native Computing Foundation Kubernetes Day India 2019, 29 Mar 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Making Cloud Native Computing Universal and Sustainable - Dee Kumar, CNCF

Description

Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io

Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects

Making Cloud Native Computing Universal and Sustainable - Dee Kumar, CNCF

About Dee Kumar
Dee Kumar brings over 15 years of cloud computing experience as the Vice President of Marketing, Developer Relations at the Cloud Native Computing Foundation(CNCF). Most recently, Dee was Director of Marketing at Docker where she built and launched developer programs. Prior to that, Dee was Director of Cloud Services Product Marketing at Equinix. Dee also held marketing leadership roles at Lenovo, Hewlett-Packard, and Dolby. Dee brings a passion for both open source and community development to the organization.

She has an MBA from Brown University and a BS in Computer Science from BIT.

A

Greetings: everyone I'm Dee Kumar, vice president of developer, marketing with a cloud of computing foundation, first I want to say I'm so thrilled to be back in Bangalore. I live in San, Francisco California, but I grew up here. So this is my town, so I'm gonna say something in color dye. Just to refresh my childhood memories, sailor Eagle, namaskara.

A

So I head up developer marketing at CNCs, so I'm a marketer by trade, I'm very much an engineer at the core, so I love working with developers around the globe, building local communities in the cloud native ecosystem.

A

Cific Google digital transformation believe it or not. You will end up with approximately 300 million plus search results and yet I'm unable to find a good definition for digital transformation, but as a part of the cloud made of computing foundation, we work very closely with analysts and we also observe the industry and based on industry reports and everything that we are hearing. There are four major factors or phenomena that are driving this change, so more buzz words here again, but I do my best to get into the specifics.

A

So the four main factors are economic ecosystems, democratization, acceleration and collaboration.

A

So, let's start with democratization, why do companies open source projects so using Google as an example and kubernetes, which is the leading orchestration platform, was a Google led project industry leaders like Craig McKee and many others? They came together to solve some major problems in the industry. How do we accelerate cloud native computing, whether it's DevOps micro services, orchestration containers?

A

How do we create a portability layer for the cloud? And, lastly, how do we achieve digital transformation? Where do we make it? Go has been using containers internally for many years, but Google alone is not able to change the broad perspective in the industry around modern applications. So there is a need for industry leaders and influencers to come together to solve the major problems that I just talked about, and the beauty of these industry leaders is.

A

They represent the collective views of all of us developers across various companies, ID practitioners and they also represent real-world end user customer use cases, and what that means is it really helps to drive the broader change in terms of how we think and think about modern applications, whether it's hybrid cloud or moving portable units to the cloud, no vendor lock-in, etc.

A

So, in order to achieve this, broader impact Google donated kubernetes to the cloud native computing foundation, so this led to the birth of CNCs in 2015 and kubernetes became the anchor tenant for CN CF CN CF was founded with 28 members and today is Dan Kahn mentioned in his opening remarks. We are, we have a 375 plus members and growing. We are the home to 34 cloud native projects and technologies. We have 83 kubernetes certified providers, including some of the largest public cloud providers, as well as large enterprise software companies.

A

As you can see from this chart here, the adoption of kubernetes continues to move to the right. I just pulled out some recent stats for you all today, so kubernetes has over 100,000 commits over 25,000 developers, 85,000 pull requests and thanks to Nikita, who pointed out to me that kubernetes has about 49,000 930 stars on github and compared to I would say: 1.5 million open-source projects on github kubernetes ranks number nine for commits number two for authors or issues.

A

Second, only to Linux and last year, kubernetes was the first CNC a project to graduate followed by Prometheus and envoy, and again this year we have two another two other projects that graduated, which is Cody, honest and continuity. The other thing that we were observing from a contribution standpoint, less than fifty percent of the contributions are coming from the u.s.. So it's more from the rest of the world like India and China.

A

Now, let's talk about collaboration. Why do how do companies maintain control when all these projects are community driven?

A

So for that we first need to understand who makes up our community, so the community is comprised of developers, contributors, maintainer program managers, project managers, community leaders, and they are the heart and soul of the cloud native ecosystem and Sarah Novotny was a good friend of mine, she's, the head of open source at Google, and she always says: let's see, folks, take more pride in shopping word and carrying water, and what that means is there are a lot of I would say non sexy tasks that need to happen, and that's why the community, regardless of the companies you work, for we collectively make things happen and that's the meaning of let's take more pride in shopping word and carrying water.

A

The other thing I, wanted to point out about the open source community is the community again demonstrates a culture of very strong core values and if the result of the community's hard work, that's left rise of kubernetes massive adoption and production established organizations like uber left, Blackrock, Bloomberg, eBay, New, York Times. All of them are now using coconut ease in production and a massive scale.

A

We work very closely with Red Bank and, according to a recent survey from red Munk, 70 percent of Fortune 100 companies are using containers and more than 50 percent of Fortune 100 companies are using kubernetes as the orchestration platform.

A

The other thing I always get asked is: is the open source community like a standards body and again, Saturn evadne put this into perspective for me. So in the 1990s, the standards body, they first started out with us 5, and then it was followed by implementation, which is programming coding, which then led to all the processes and governance around that.

A

But the open source community approach is more like better, so you start with code first and that and then equally start into a spec which then becomes the standard again, as telecon mentioned, we are very grateful and thankful that our end user community is growing. We have 79 companies in our end. User community also want to take the time to define what do we mean by end users.

A

So end, users are essentially really great brand companies where they are using and trying cloud native technologies internally, and they are not commercializing or selling cloud native technologies externally, and so we work very closely with their user community and it's really important because they are like the early adopters and they're taking big risks with trying out all the open source projects and what we do is we make sure we gather their fee. There are requirements and which, in turn is helping shape.

A

All the cloud native projects hosted on the CNC F I'm, making these projects more reliable now also wanna touch upon. There are knots of always not all open-source projects of success. Successful, so I wanna share a little story about this company called cluster HQ. So few years ago, cluster HQ took about 80 million dollars in venture capital funding and then the company shut down, and then there was a blog post from the CEO which read be a cluster bomb.

A

So what went wrong so the customers of cluster HQ were definitely not hot, so there is a strong need for the infrastructure software to support state food containers. So flogger was an open source project from cluster HQ, but the issue was slaughter was connecting to a wrong storage. So what I mean by that is, it was connecting to legacy storage, architectures, so essentially, storage, arrays that were meant to serve storage for service and not for applications running in containers.

A

So, though, cluster HQ set the path for the industry to think about infrastructure software to support stateful containers where cluster HQ went wrong is they came to market pretty early and then also at that point in time the legacy storage architectures were still evolving and today Phil knows: if you see the legacy, storage architectures have evolved a lot, thereby supporting stateful containers.

A

It's also really important and again I think Diane touched upon this slide in his opening remarks. Is we need to understand what do we mean by crossing the chasm? So there is a book by Jeffrey Moore, where he talks about five segments of Technology consumers. So the first segment is innovators. These are techies who live on leading edge and they love to play with new technologies and make these new technologies work.

A

And then you have the earlier adopters, where essentially visionaries and what they do is they try to take the work from the innovators and push it through large organizations and they actually take the risk to use these technologies. New technologies against highly visible high risk projects, and then you have the early majority who are pragmatists and all they care about is to improve productivity with existing operations, and hence lies the chasm so the first two segments they want to.

A

They want change and disruption, and the early majority of the fog matters they're just looking for continuity, and it's not, coincidentally, that the stages of projects within the CNC F neatly mapped the segment's. So we have three stages for our CNC F projects. It's a sandbox incubation and graduation, and then there is one third of the end users who are the late majority, and they just want things to work.

A

They don't want to spend a lot of time and invest a lot of time and deep expertise and offering a managed service like kubernetes, could really help them with ease of setup, as well as to reduce operational complexity, so as technology becomes more and more reliable, that allows for the late majority to adapt that again allows for the influx of new wave of sidewalks sandbox projects, thereby resulting in more innovation.

A

The next thing I want to talk about is economic ecosystems really want to thank all the members AWS Microsoft, a number of Platinum members here who are really committed to open and the cause of open source, as well as the community that is growing around it also wanna thank our some of the board members, all of our members, jay, frog, digitalocean, etc, for really again being a part of this community and believing in the cause of open source and again it's because of the support that we are getting from these members.

A

It is resulting in flourishing ecosystems.

A

So thanks to Dan kahan, who is the creator of the landscape, and then the excitement is actually seen in the cnc of landscape, where, as we have all these members who are coming in and becoming a part of the foundation as a concealer, there is a challenge. The logos are becoming really really small, but I. Think it's a great problem for us to have, and what that is. What that is representing our representative is there is a lot of. There is an increase adoption of cloud native projects for technologies.

A

Sakuya, ladies, has developed and grown to becoming an economic powerhouse, thereby loving number of projects to be built around it, and what is that has done is its resulted in opportunities for companies of any size to start creating products and services, and thereby that again results in a cycle of flourishing economic ecosystems. I love this life. This is from Jim's Emlyn who's, our executive director for the Linux Foundation, and he really thinks about how do we sustain this innovation?

A

So you have the projects and then those projects are being used to create products and then every one end of the day, things of our profits. So the way we want to think about it is our jobs here collectively is to make sure that these three gears are really spinning faster and our job in C&C F is to make sure that we host really good open-source projects.

A

So all of you and all the different companies you're a part of- can use the open source projects, and then you build products and services on top of it, so you can commercialize it, thereby resulting in value which is profits, and then the beauty of this whole cycle is you, then take the profits and you reinvest it back into the open source projects and what that means is you hire more engineers?

A

Developers means more code, more improvements and thereby resulting in more products and services, and then it just becomes a virtuous cycle of investment, innovation and economic activity. I.

A

Guess you know all effort from all the thoughts since morning the path to cloud native can be pretty daunting, and so one of the things we have done isn't in CNC FS we've created this tool, which is the cloud native trail map, and essentially this is a I would say a get started guide. It can be more like a pocket guide if you're just thinking about its transformation. Where do I begin again, it's not like in your specific order. There are innumerable paths to get on this journey.

A

It can be very specific to a different use case. You can perhaps have step one and jump to another set of steps in the trail map. So the way we came up with the trail map is to kind of think about. When you're going on a hike, you essentially have a map, and then there are different trails to pursue.

A

So we use that as an analogy to kind of simplify this whole journey to cloud native and also having talked to the end user community I can say that this path is kind of more of well traveled, well tested and well it rusted.

A

So there are like 10 steps that we define as a part of the cloud native trail map, so quickly want to go through the 10 steps.

A

So the step one is contained, so you can do cloud native unless you containerize your application, so it doesn't matter what size application is any size application will do, and even if you have a pdp-11 old code running in an emulator, you can still can in our eyes that Ticketmaster is a good example where they did have pdp-11 code running and then they were very successful in container izing their application.

A

The second step is, or often there's some confusion where, once you containerize people immediately think that you need to orchestrate it, but then there is a set of processes. You do need to set up continuous integration and continuous delivery. So the changes that you make to your source code results in containers that are automatically built, tested and deployed across various environments, whether is develop a Jane moving to QA and to production. The next thing you want to think about is to set up automated rollouts rollbacks, as well as testing CN CF.

A

We don't offer a CI CD solution, but there are some solutions to think about which is Jenkins, J, frog and spinnaker. We did announce a new foundation called the CDF at the Leadership Summit couple weeks ago, where we're primarily going to be focused on the CSE D lifecycle and in the projects an ecosystem around that step. Three is orchestration, so pick an orchestration solution.

A

Kubernetes is definitely the leading orchestration platform. There's a certified version, that's available, but you can also think of other orchestration options that are available, which is start with suam resource pouch, pneumatic cetera, step. Four is observability. Observability is essentially table stakes for cloud native at scale, so you want to then think about monitoring logging as well as tracing for monitoring. You can consider Prometheus for logging flow NT and for tracing.

A

You want to look for an open tree compliant solution like Yaeger step 5, a service mesh I think we all know service mesh by its name says that it's all about connecting services, discovery of services, health check, checking routing and it's to use for monitoring ingress from the internet. Some of the solutions to consider here are envoy, liquidy and core DNS step. Six is networking, so it's really important to enable a more flexible networking layers.

A

So it's important to look at a C&I compliant network project like calico we've met or flannel distributed database again, if you're looking for more resiliency and scalability, you will not get that with a single database, so you want to look at Vitus, which is a distributed database, and it's really a good option for running my sequel at scale through sharding step. Eight is messaging so again here when you're looking for higher performance than json dressed g RPC is a good solution to look at nats is a messaging system. Again, it's more middleware oriented.

A

It has a publish and a subscribe kind of messaging system, and it also is applicable and takes care of many newer use cases like iot container around time. You can use alternative container on times, I think it's important to look at it over CI compliant version like container D rocket or cRIO alibaba's pouch container. Our time is also a good option.

A

The last step is software distribution. Again, if you need to do a secure software distribution, you might want to evaluate notary which is based on the update framework, which is called tough, so tough, again is defined for with really large part of a large software distribution framework and the beauty of tough is it's very secure and the status it incorporates cryptographic keys, which again has different kinds of keys for content, verification, content, signing and then the tuff framework is designed to more or less address a lot of their different use.

A

Cases in terms of security attacks and tuff is being used in production by darker VMware leap and other companies that that are listed here. Harbour is again a secure container registry, and now it's a part of a CN CF project.

A

So the last factor we want to talk about is acceleration, so this is really good news we again and C in C F. We do a survey each year. We do it three times in English as well as in Chinese. As you can say, some amazing results here, based on the survey results. When we talk to enterprise companies, may we really focus on developers and IT practitioners? We are seeing that cloud native technologies in production has grown over 200% and some of the main benefits that folks are seeing in this cloud.

A

Native ecosystem is primarily around faster deployment. Time improves, scalability and, of course, cloud portability.

A

Lastly, I wanted to just share a story from Pinterest. Most of us have interests, have been in existence for I would say eight years now and they've grown so fast, so that is resulted in about thousands of micro services and a lot of infrastructure layers that are old, so Pinterest went on a mission in terms of wanting to launched a new compute platform very quickly from idea to production.

A

So they went about doing that by putting all the thousand micro services inside a talker and once they did that part of phase two, they started evaluating orchestration solutions and then they settled on kubernetes, and one of the use case they carried you wanted to use is to make sure that the Jenkin workloads, which is essentially serving the Pinterest website.

A

It was a static cluster, but then they move that to the kubernetes environment and what they saw was some amazing results that I wanted to touch on is they saw a they were able to reclaim 80% of the capacity during non-peak hours, and then they also saw a reduction in 30% in terms of the instant instant hours per day for use in terms of serving contents for the website.

A

And again, we focus on our end user community, and we have some amazing case studies so feel free to go to our website CLC after I/o, where you can get a lot of great stories in terms of how our end-users are trying and experimenting and using cloud era technologies. Thank you.