AT&T, CoreOS, Weaveworks, and Fujitsu Share Thoughts on Cloud Native Market

By June 28, 2016 January 15th, 2018 Blog

Some of our member companies recently shared the market problems and misconceptions around cloud native, as well as the objectives and benefits of CNCF.

Here is what AT&T’s Lead Principal Technical Architect, Domain 2.0 Architecture and Engineering, Cloud Architecture, Doug Nassaur; CoreOS CEO, Alex Polvi; Fujitsu Director of Development Department, Linux Development Division, Kenji Kaneshige; Weaveworks CEO, Alexis Richardson, had to say.

1. What are the cloud native market problems?

Polvi, CoreOS: We’re at the beginning of the adoption curve in cloud native computing. Naturally, new markets have a growth period before stabilizing. Any emerging market and ecosystem needs to have hobbyists, commercial vendors, and vendor-neutral industry groups to grow. The CNCF is the latter for this ecosystem.

Kaneshige, Fujitsu: Cloud native is a set of technologies, but to get the benefit from “high velocity”, users need to understand what culture and rules. Organization is required for delivering apps and change, so there will be some barriers. Providing case studies that show the benefits to the market will be helpful.

Richardson, Weaveworks: Customers want to build cloud native applications and microservices.  But many are not sure where to start, and what tools to use.  We need to help by identifying the best tools, patterns and practices, and then explaining these clearly.  There’s multiple issues like data, networks, monitoring, security, governance, and orchestration. On top of all that, companies want to deliver apps with high velocity.  The CNCF can show the way to do this.

2. What are the biggest misconception about cloud native in the market today?

Nassaur, AT&T: We have an impedance mismatch between expectations and effort. Too many people believe containers to be the next silver bullet – put your software in a container and it magically removes all barriers and obstacles to performance, availability and economics. You can’t send your application to the gym each morning, have it run two miles every night, put it on a gluten-free diet and call it a micro-service. It doesn’t work that way. Cloud native applications are container-packaged, dynamically scheduled and micro-services oriented. This means they are supported and enabled by a software ecosystem which must be in place and based on open standards – that’s why we need the CNCF.

Polvi, CoreOS: There are some misconceptions to address head on and we are focused on addressing this in the CNCF. One misconception is that containers will replace virtual machines. This is not true. For example, many container deployments are on AWS, AWS uses virtualization for every server, and I do not see that changing for the foreseeable future.

Another myth is about the security of this style of infrastructure that may have come from faulty early implementations. This style of infrastructure is actually more secure. With the ability to isolate individual applications and automatically update the software stack, containers are a great way to increase security.

A misconception is that this style of infrastructure is only for new applications and that old enterprise apps can’t run on cloud native infrastructure. This is in fact how the biggest companies like Google run all of their infrastructure, and have been for years given the balance of operational efficiency and flexibility. These benefits in addition to the ability to have consistent environments will drive the adoption of cloud-native infrastructure over the next few years.

Richardson, Weaveworks: The biggest misconception is that you need to understand infrastructure technology in order to deliver cloud native applications. Fundamentally, cloud native is about enabling application developers to create business value, without thinking about infrastructure.  And the second big misconception is that PaaS delivers all things cloud native.   That is certainly not true – PaaS describes an opinionated model for application delivery, and is just one of several cloud native patterns.

3. How is CNCF uniquely positioned to address these market problems and clear up these misconceptions?

Nassaur, AT&T: Integration, interoperability with freedom of choice to adopt and consume innovation as each building block matures without having to turn over or upgrade the whole farm. Encouraging and supporting new projects is vital for innovation. Education, awareness and being an impartial lighthouse to guide newcomers will expedite understanding, and as a result promote adoption.

Polvi, CoreOS: As a vendor neutral home for cloud native computing, CNCF is positioned to help unify and clarify what is happening in this market.

Kaneshige, Fujitsu: We’ve got a lot of amazing companies involved. If you look at the governing board and technical oversight committee, it really is a who’s who in the industry right now. Lots of technology industry leadership collaborating.

Richardson, Weaveworks: The CNCF is pragmatic — we understand that customers want to adopt cloud native incrementally, and may not want to replatform or make wholesale architectural changes.  We’re not locked into a particular architecture. We’re trying to align with users’ needs as they emerge. We have a model that is designed to grow with their needs. We’re going to listen to users, which is why we created an end user board.

4. What are the benefits CNCF aims to provider developers, end users, and CIOs?

Polvi, CoreOS: For developers, the question should be how can CNCF help open source projects make cloud native possible? For instance, with Kubernetes, being vendor neutral from a copyright perspective helps a diverse set of companies get involved with the open source collaboration side of things. The CNCF neutralizes party alignment from the different vendors.

Kaneshige, Fujitsu: For end users, it is very important to provide a stable and open cloud native technology and accessible APIs. For this purpose, the cloud native technology should be harmonized by the industry leaders. CNCF has a very good opportunity to do that.

Richardson, Weaveworks: CIOs want IT to go faster, software to go faster. CNCF will provide them with the ability to go faster. They will be able to adopt a set of technologies and practices that will enable them to create software in-house for business people working with IT people to keep up with companies like Airbnb, while delivering better services to their customers. Therefore, the obvious benefit to end users is better services.

5. What are some of the key objectives for CNCF?

Nassaur, AT&T: To provide education and awareness. Cloud native is not a single architecture. Customers have multiple use cases that have quite different approaches, so it is important to support multiple, different projects. We need to educate people on the need to invest in a cloud native fabric by showing them the benefits of all kinds of open source componentry as it sits today. As it continues to evolve, they can get the benefit of the excellent innovation from the open source community, while still having a consistent, predictable, repeatable infrastructure that they can count on to run their organizations.

Polvi, CoreOS: The main objectives we see are:

  • Lowering the confusion of what’s going on in this space.

  • Providing enterprises an adoption path.

  • Fostering, supporting, and generally helping the open source projects that make cloud native computing possible in the first place.

Kaneshige, Fujitsu: To make cloud native architecture available, we would like to grow the ecosystem with more projects and to establish a set of standard APIs and software of cloud native technologies for collaborative computing.

Richardson, Weaveworks: Reduce confusion in the market and make it faster and easier to adopt cloud native applications to help customers make that transformation faster.