Snapt, an application delivery controller (ADC) company, recently announced the general availability of Snapt Nova, a centrally managed, any platform Load Balancer/ADC management and app visibility platform, which includes a Load Balancer, Web Accelerator, Web Application Firewall and GSLB.  To find out more, VMblog spoke with Dave Blakey, co-founder and CEO of Snapt.

VMblog:  Snapt just recently announced the launch of the Nova ADC platform, tell me what impact you expect Nova to have on the market.

Dave Blakey:  The number of endpoints and the amount of data going to the endpoints in application delivery is exploding. More servers are serving more data to more clients, from more locations. This is going to continue to grow exponentially, made more difficult  by the need to deliver even faster than before.

These modern DevOps and platform-centric designs are going to grow in complexity and popularity. The increase in multi-platform, multi-location delivery of applications and the addition of service mesh ADC will drive a massive uptick in the number of ADC (v2) endpoints.

VMblog:  How does Nova represent a sea change in how DevOps and how IT Ops teams manage their networks and application availability?

Blakey:  We know that with the four pillars of ADC guiding us-- security, availability, performance and observability--we have developed a highly scalable, artificially intelligent, automated solution for availability, security, performance and observability in any platform - from hardware to cloud, containers to service mesh. 

VMblog:  How does Nova address the unique challenges arising from cloud native, microservices (east-west) architectures? How is Nova's core philosophy part of the solution to these challenges?

Blakey:  Snapt embraces north-south and east-west application architectures, regardless of their platforms. This enables traditional IT operations and DevOps to be agile while reducing risk.  Our core philosophy is fourfold - ADC should be a service and not a server, the control-plane / data-plane architecture is the only way to manage ADC at scale, ADC needs to be in VMs, Clouds, at the Edge, in containers, service mesh, and everything in between and the network must drive itself - using AI, ML and automation. 

VMblog:  What are the benefits of Nova's data plane centric approach?

Blakey:  To address the explosion of services, and the changing nature of where they exist, a true control-plane / data-plane architecture is required for managing ADC endpoints. Specifically, this means that your control-plane (central management) is able to do the following things:  It must be in constant communication with all your worker nodes.  It must be able to deploy new nodes, self-repair nodes and shut down unnecessary nodes.  It must be able to store all your configuration files, rulesets, etc. and centrally to deploy anywhere, as well as entirely recreate a worker node with no human interaction to scale or repair.  Finally, it  must store and provide metric and telemetry information for all workers and provide a global on-platform GSLB for routing traffic to changing nodes. 

This design ensures that you have no value in the data-plane, that worker ADCs or nodes have no state, and that you can scale without limitation due to controlling the configurations and routing to worker clusters.

VMblog:  What challenges are facing the ADC market today?

Blakey:  Platforms are on a spectrum, they are not a destination, so an ADC provider needs to support running on any platform, to the degree that you are available as a vendor-neutral container or binary that can run on any modern Linux platform.

A service mesh load balancer means you need another vendor for cloud, and another for on-prem. ADCs need to act as the ingress (and mesh) for VMs, clouds, Edge compute, containers, service mesh and the balance.  This creates several technical requirements:  ADC workers must be extremely lightweight, able to deliver high performance on edge and light compute containers and they must be truly platform and vendor neutral. The most practical example of this is the ability to run on any Linux system, allowing clients to deploy the solution to every location they have applications.

ADC workers must also be able to scale-out or scale-up to handle load in varying architectures, e.g. tin vs. containers (more cores vs. more pods).    

VMblog:  Tell us about Nova's pricing structure and how it's different from the traditional ADC pricing models?

Blakey:  We set up Nova's pricing structure to be as flexible as possible so no one pays for something they don't need.  It's also on-demand, consumption-based and multi-platform.  For example, if a customer pays $99 a month for a single node (Nodes can run any number of ADCs, a Node is typically a container, VM, cloud instance, or server) and runs it for five hours in total during a month, they will be charged 67 cents. This means your ADC is billable on the same terms as your cloud storage and compute resources. 

Once you start paying for a Node, your account becomes a Paid account, and all Nodes are charged for as they all receive additional functionality, storage, etc.  We also offer Nova Community Edition, which is free and allows up to five nodes!  You can use and deploy as many ADCs as needed as a community user.

VMblog:  What's next for the future of Snapt?

Blakey:  Our future investment is focused on three main areas: scale, east-west application design and service mesh.  We think about scale through the lens of "just-in-time" - using predictive analytics and machine learning we scale customers in and out within 60 seconds of the opportunity.

Cloud-native's real meaning, and real goal, is to abstract the data plane from the control plane and allow applications to be vendor, location and platform neutral and capable of running any number of systems as the load requires and this is our vision.

Nova already supports service discovery and ingress traffic in service meshes and very soon will support a full mesh with ingress and egress between nodes inside the mesh.   We are continuing to develop heavily in this direction based on our belief that all applications regardless of architecture need the four pillars of ADC.

##