The Focus 4 Cloud Native in 2019

Joseph Cooper
KintoHub
Published in
4 min readApr 24, 2019

In the last 18 months, I’ve spoken from everyone to people who have played with microservices + cloud-native apps, people who know about them and people that ain’t got a clue what’s going on. Being in any three of those categories is okay — zero judgment here folks.

I’ve spoken to over 400 companies and developers in 2018 and about 50% are just starting to do proof of concepts or migrate over to Kubernetes. The “hype” is real as they call it, but it’s been a multi-year journey for cloud native apps. K8s 1.0 Launched in 2015 and is now the most contributed open source project the world has seen.

Now to answer everyone’s question if you don’t know why this topic is important. Google, Amazon, Microsoft, Facebook and every other giant that pumps billions of dollars of revenue into their pipes have invested billions of dollars into tech and infrastructure related technologies. Some of these technologies have been open sourced for the greater good, but the average team doesn’t have the budget to use them properly. It’s quite ironic, but fortunately not a permanent dilema.

KintoHub’s mission is to give average engineering teams the power to do what the big boys and girls do - Rapidly Release @ Massive Scale. Smart rapid releases supercharge your product to fail fast, iterate faster & aid your chance at making success a reality.

The power of the cloud on your machine?

KintoHub has had the honor to have a lot of fun over the last 15 months playing with all the tools that the open source world has to offer. A bit too much fun, but enough to find our focus on what we believe people need to release faster and stronger.

Unfortunately, only the people who have played with cloud native apps and microservices, for at least 3 months full time, are the only ones that really know the pain. The press & open source world does not share this pain often, but it’s real, and it’s BIG... There are too many options in the cloud native world, that you can lose months of time trying to choose the right one and make it production ready. And that's 1 out of at least 10 you need to glue together.

And with that said, that’s what KintoHub now here to solve

Although that we’re making it easier to get started, there are setill a lot of open ended questions in the cloud native world that need to be explored further. Here are the top ones that we should focus on in 2019

  1. When do I use serverless, what if one lambda function depends on another lamda function which then depends on a serverless Aurora service? The dependency chain scares me alongside from the 1–3-second cold starts. (Proposed Solution): We should pre-warm dependencies when API calls come in. This could be done with HTTP Options calls fired to all dependencies when the head of the dependency chain is called.
  2. CNCF has 100’s of options for me to choose to build my microservice platform. And only a few have been graduated so far. What do I choose for my service discovery, encrypted secrets, CI/CD workflows, message passing + queues, metrics, log storage, alerts, analytics, ect? (Proposed Solution): Teams should open source their “core” infra which combines these projects. KintoHub is considering this soon.
  3. Developers keep pushing releases to DevOps but don’t care to debug issues with their apps due to lack of tooling to get logs and metrics about their services. What Should I Do? (Proposed Solution): Every developer should deploy a container into k8s. It takes a couple hours to setup a managed cluster and do this & it feels magical. Inspire developers to become engineers and hire problem solvers not only experts in a single domain! Aside from this, get jaeger & kibana setup in your environment to aid diving in deep. Or just use KintoHub :).
  4. How do I share session information across several services in my microservice ecosystem? One DB per Microservice, why does this rule not occur for shared memory as well? (Proposed Solution): KintoHub is experimenting with Session Memory and Global Memory which get/set data at the proxy layer. It works well in controlling access to who manages the data. But open to hearing more solutions around this!
  5. What’s the best way to manage my cloud native app and/or potentially 1000’s of lambda functions? (Proposed Solution): KintoHub is working on this, but www.serverless.com is doing a great job as at managing functions too!

I’m excited to see more sharing in 2019 in the cloud native world. We have the tools at our finger tips, but how to get them working at billion user scale is still too challenge. Once we solve this together, teams can finally build truly infinitely scalable products without an ounce of thought towards infrastructure. We’re in an amazing era of abstracting the cloud itself!

--

--