Derek Collison is the Founder & CEO of Synadia. Synadia invented the open source connective technology, NATS.io. With NATS you can easily connect all your applications and data at a global scale, no matter what language they are written in or where they are running.
Derek Collison is a 30 year industry veteran, entrepreneur, and pioneer in secure and large-scale distributed systems and cloud computing. He helped change the way financial, transportation, and logistics systems fundamentally worked while spending over a decade at TIBCO, designing systems that still power much of those industries today. At Google, Derek co-founded the AJAX APIs group and created the largest CDN for popular javascript libraries, identifying the need for easy and secure access to Google’s services without the need for additional servers.
How can AI and edge computing work together to achieve low-latency, complex data processing?
Edge computing has two primary roles to play in the AI economy. The primary role today is in the collection and distribution of edge based data / telemetry for use in supporting the centralized training of models.
The emerging and much larger opportunity is for AI inferencing use cases. AI and edge computing are a critical and powerful combination that enables real-time processing, reduced latency, improved accuracy, and enhanced privacy & security.
We are now seeing clear signs that inference at the edge is not simply getting large quantized models to run at the edge. It’s evolving to a multi-stage pipeline, one where prompt augmentation and retrieval-augmented generation (RAG) operations transform the initial prompt from a user or system using real time and location independent access to data at the edge. The second is a graph traversal of multiple models vs. a single model whether they are local or remote.
What are some real-world applications where the combination of AI and edge computing has made a significant impact?
AI is impacting every industry at an increasing velocity. We are seeing the open-source NATS community and ecosystem building applications and services in industrial manufacturing and IoT (predictive maintenance, process optimization & digital twins), autonomous vehicles (optimized routes & safety), green energy (energy optimization), retail (personalized shopping experiences and real-time inventory management), FinTech (regulatory compliance and algorithmic trading) and government (DoD/IC communities e.g. drone CUAS).
One of our customers is a manufacturer of medical devices that are used as a training platform at facilities all over the globe. Each local event needs to be captured and stored as part of a certification process. Having to trombone all traffic back to the central cloud was both cost prohibitive and difficult for the more remote sessions. Our native edge capabilities allowed them to simplify their platform delivery capabilities.
As the creators of NATS.io, how do you see open-source technology shaping the future of edge computing?
Edge computing has proven itself to be a different landscape than cloud computing. The approaches and solutions used in the current so-called “cloud native” architectures are not translating to the edge.
The “what” component of the solution remains familiar, but the “how” is very different. Synadia/NATS.io changes the “how,” not the “what,” to solve challenges with micro-services and data access and flow at the edge across any cloud or geography.
What are the key benefits of using NATS.io for global connectivity of applications and data?
The main advantages of NATS.io come from its unique approach to connectivity and integrated data layer. NATS’ intelligent connectivity enables applications and services to be built with location transparency or what we like to call “nomadic applications.” NATS removes the requirements of extra cloud infrastructure such as load balancers, service mesh, or API gateways to allow users to write an application once and deploy it anywhere.
Building upon that is our data layer, with streams and key-value and object stores. It’s the combination of our data layer with the connectivity layer that gives powerful access and flow patterns that are suited very well to the edge for latency sensitive and real time applications.
What are the next big innovations in edge and connectivity that you foresee in the coming years?
Real-time fleet management, telemetry data collection, analysis at the edge and of course AI inference at the edge in real time with specific temporal envelopes will all be priority innovations. Examples of this include:
- Increased AI capabilities at the edge: As processing power improves, more advanced AI and machine learning capabilities will move from the cloud to edge devices.
- 5G and edge computing convergence: The continued rollout of 5G networks will transform edge computing capabilities by providing ultra-low latency and high bandwidth.
- Edge-to-cloud continuum: We’ll likely see a more seamless integration between edge and cloud computing, creating a continuum where data and workloads are dynamically deployed based on requirements. We like the term “truly nomadic applications.”
- Hyperlocal data centers: There will be a proliferation of small, hyperlocal data centers to support edge computing needs. Companies like Akamai are already repurposing infrastructure to create a global network of edge data centers.
- Industry-specific edge solutions: We’ll see more tailored edge computing solutions for specific industries like healthcare, manufacturing, retail and agriculture addressing unique challenges and opportunities in each sector.
How is Synadia positioning itself to lead in these future trends?
Our approach is first to tackle network connectivity, then data and we recently introduced the ability to handle workloads natively with the NATS Execution Engine (NEX). This differing approach from others caused us to be very misunderstood early on. However, this approach has allowed us to bring some innovative and very powerful solutions to market challenges around system design and edge.
As applications transition to microservices and move to the edge, we need to ensure the various components can communicate. We flip the normal time-consuming approach of adding API gateways and service meshes, and instead leverage Synadia Platform and NATS.io to simplify the tech stack and its communications.
What are the main challenges in designing and building modern distributed systems, and how does Synadia address these challenges?
Most folks design their systems with a focus on workloads first, then data, then finally the network needed to put those together. Security is usually layered last.
Synadia’s tech stack, anchored by NATS.io, allows consistent and rapid design of modern distributed systems. These can adapt to single cloud deployments, but more importantly remain future-proof and secure when multiple cloud providers, regions and edge locations enter into the picture.
Can you provide examples of how Synadia’s solutions have helped companies overcome issues with traditional connectivity models?
Synadia’s platform allows system designers to achieve rapid success into edge-based deployments and to future proof those systems regardless of deployment targets.
They also allow a consistent security model and the elimination of a bunch of other moving parts not needed with the Synadia platform, like load balancers, api gateways, GSLBs, firewalls, VPNs, service meshes, etc.
We work with a global auto OEM that wanted to modernize its digital infrastructure. The goal was to avoid siloed application projects and deploy a “global data fabric” that all of its future digital services could plug into. These are the types of projects that excite us.
How do Synadia’s commercial offerings, such as Synadia Cloud and Synadia Platform, enhance the capabilities of NATS.io?
While we will continue to open source features targeted for a NATS developer, Synadia provides value added IP primarily around the enterprise monitoring and administration of NATS.
Synadia Cloud & Platform are “certified” versions of the software that are supported for mission critical and enterprise deployments. These are packaged offerings that provide a spectrum of offerings from lower cost, multi-regional and multi-cloud SaaS from Synadia Cloud, to on-premise and edge location deployments for Synadia Platform. These two often work together as well.
How important is community engagement for Synadia, and what initiatives are you taking to foster a strong developer community?
Our technology roadmap is completely driven by our OSS ecosystem and the developers using the software. We are constantly engaging with the ecosystem, interacting and getting feedback.
We have made a significant investment in resources to help with the developer experience. This includes documentation, video tutorials and sample code resources.
Thank you for the great interview, readers who wish to learn more should visit Synadia or NATS.io.
