Virtualization Technology News and Information
VMblog Expert Interview: Krisztian Flautner Talks Banzai Cloud, Kubernetes and Hybrid Multi-cloud Solutions


To avoid being tied to one powerful cloud provider, companies are becoming increasingly interested in hybrid multi-cloud solutions, opting for adopting a combination of cloud technologies purchased from big cloud suppliers, and some computing in-house.  Although hybrid cloud is a great operating model for enterprises, it has been very difficult to pull off.

VMblog spoke with Krisztian Flautner, CEO of Banzai Cloud, to get his perspective around this topic. Among other things, he shares his expertise on how Kubernetes and cloud-native technologies open the door to convergence between the technologies that are used in the data center in enterprise settings and those that are driven by the clouds.

VMblog:  What kinds of cloud solutions are enterprises moving towards and why?

Krisztian Flautner:  The cloud provides an attractive value proposition for most enterprises but reconciling the technologies and operating models between classical data center-oriented IT systems and cloud-centric approaches has been a drag on adoption. Kubernetes and many other Cloud Native technologies have opened the door to convergence between enterprise IT and the cloud but this path has its difficulties as well; different architectural patterns, skills shortage, access to know-how, complexity have been creating headaches and slowing adoption. The solutions that help enterprises get there need to simplify the problem and bridge the gap between enterprise- and cloud-oriented thought patterns.

VMblog:  What makes hybrid multi-cloud solutions difficult for enterprises to pull off?

Flautner:  There are two main groups of issues: a) application portability and b) operational consistency across clouds and data centers. Application portability is not just about eliminating dependencies on cloud-specific components but may involve re-architecting the application's data model for changed circumstances. Not all parts of an application can (or should) be moved from one cloud to another, it might be desirable for some of its components (e.g. a database) to stay in one place regardless of where the rest of the application is run. These changes may drive new architectural requirements into the application so that it can deal with increased latencies, bandwidth and consistency issues better. In our experience, success with hybrid multi-cloud deployments is a strong function of the attention to detail (and quantification) on the data side. Operational consistency - in many ways - is a tougher nut to crack. The standard operating model is that all clouds provide their own solutions for the various layers of the technology stack (e.g. security, logging, monitoring, etc.), which then have to be brought in synch, or replaced with, a cloud-agnostic solution for hybrid multi-cloud use-cases. The complexity of the problem increases further if we also consider interoperability with on-prem enterprise systems and the different approaches they entail.

VMblog:  How are Kubernetes and the cloud-native technologies opening the door to convergence between the technologies that are used in the data center in enterprise settings and those that are driven by the clouds?

Flautner:  Kubernetes and the technologies in the Cloud Native Computing Foundation open up the possibility of a cloud-neutral, broadly supported technology stack that can be tailored to specific enterprise use-cases. This opens up the potential for enterprise customers to better control the pace of the evolution of their technology stack and not be completely beholden to the whims of their cloud providers. This stack also enables them to unify the technologies they use on-prem, in their own data centers and in their clouds. The challenge is that the technology in the CNCF isn't a single integrated stack, its components need to be integrated and configured for specific use-cases in mind, which in turn requires know-how and testing not just of the components themselves but also the various options and alternatives.

VMblog:  How does this potential unification allow companies to start thinking about CapEx and OpEx in a different way?

Flautner:  There are multiple angles on this. If an enterprise workload is easily movable from one data center or cloud to another, then it opens up the possibility of making finer-grained trade-offs on allocation, which in turn can lead to significant cost savings. An attractive feature of the cloud is that its resources can be ramped up and down quickly, on demand, and only incur costs on a "pay-as-you-go" basis. But just as with mobile phone pay-as-you-go plans, much cheaper deals can be had if a user commits to a monthly spend, even better ones if they commit to a prepayment. But these are only good deals if one actually uses all the capacity included in the plans. The hybrid multi-cloud approach opens up the possibility for enterprises to optimize for the "base load" (the capacity they need on a long-term basis) and the variable load differently. The company may choose whether it's most cost efficient to run its own data centers or have long-term commitments with cloud providers for the base load. On the variable load side, it can choose the cheapest provider at the point of time they need the extra capacity.

VMblog:  What does Banzai Cloud do, and how does its solutions help enterprises in the changing cloud landscape?

Flautner:  Banzai Cloud's main product is the Pipeline platform, which allows enterprises to develop, deploy and scale container-based applications productively in uni-, multi- and hybrid-cloud configurations. The platform seamlessly supports five cloud providers and on-prem data center infrastructure using Banzai Cloud's own CNCF certified Kubernetes distribution. The Pipeline platform has integrated support for day 1 and day 2 operations, including a cloud-agnostic security layer, logging, monitoring, disaster recovery and includes the ability to deploy a multi-cluster service mesh across clouds. The platform's real-time hybrid orchestration and federation capability is the key that unlocks the power of the multi-cloud. This approach lets resources to be scaled up and down dynamically in response to real-time demand and involves a lot of automation under the hood. The use of the platform supercharges the productivity and flexibility of enterprise operations and enables many deployment options (e.g. quickly changing cloud providers) that had been considered to be difficult and time consuming before.


Published Thursday, September 05, 2019 7:32 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<September 2019>