Back in 2015, a team of experienced entrepreneurs looking to usher in the next generation of smart, learning-based applications at the edge founded SWIM.AI. SWIM.AI’s software delivers rapid business and operational services by identifying relevant data and learning new insights in real time, with minimal configuration or management requirements. It is designed to run local at the edge on existing equipment. SWIM.AI has delivered powerful new insights to enterprises, manufacturing environments, utilities, cities, and IoT/IIoT environments using small footprint software that ingests data, self-trains, analyzes data and learns new insights – on the fly.
Path Breaking Products
Organizations are drowning in real time data from equipment, assets, and other data sources. Hidden insights have the potential to optimize processes – but finding those insights remain a challenge. Complex, big-data ML solutions are expensive, slow and unsuited to real-time data. It is important to cost-effectively analyze and learn on data at the “edge” as it is produced. SWIM.AI has developed a powerful new way to transform streaming data into big insights – entirely at the edge, where data originates – using the compute on existing edge devices.
SWIM.AI’s software product, SWIM EDX, analyzes streaming data and learns on-the-fly to deliver insights and predictions in real-time, avoiding the latency, cost and complexity of bigdata batch analytics. SWIM EDX can be deployed on edge devices to find critical events and hidden patterns, delivering insights and predicting future behavior to inform real-time decisions. SWIM EDX automatically generates views of analyzed data from any device and delivers insights/predictions via streaming APIs for use by enterprise applications and other users.
The SWIM EDX solution rests on three key innovations:
- Edge Computing: SWIM EDX efficiently processes data where it’s created, on the device itself, to interpret and react to data without waiting for centralized processing to occur in the cloud.
- Digital Twins: Used in application frameworks such as Akka, Erlang and Microsoft Orleans, SWIM EDX utilizes the distributed actor model which represents real-world entities as actors or digital twins. The distributed actor model provides a framework that allows digital twins to learn – on their own real-world data – to predict future performance.
- Machine Learning: Learning at the edge on real-time streamed data ML approach. SWIM EDX’s ML is cost effective, real-time and can be deployed on in-production hardware – even non-traditional systems such as ARM CPUs.
These innovations can enable the creation of a variety of real-time applications, including predictive maintenance, infrastructure monitoring, autonomous vehicles, AR/VR, gaming, manufacturing, Smart Cities, and many others.
Rusty Cumpston is the CEO and Cofounder of SWIM.AI. Rusty has an established record of startup success having co-founded Sensity Systems, Inc., an intelligent lighting and smart city software provider, prior to its acquisition by Verizon Communications in 2016. Rusty has also served as Chief Executive Officer of CloudShield Technologies. Prior to CloudShield, he served as the Chief Operating Officer of ONI Systems, a leading provider of optical networking systems for metro transport markets. Rusty was the Interim Chief Executive Officer and Vice President of Engineering at XenSource and the Vice President of Engineering at Infinera (IPO 2007). He graduated from the University of North Carolina at Chapel Hill.
Chris Sachs is the Chief Architect and Co-founder of SWIM.AI. Chris has developed distributed real-time solutions for some of the most challenging data use cases around the world. Applications that Chris has worked on continue to power airports, malls, intelligent traffic systems, manufacturing systems, and more. Prior to founding SWIM.AI, Chris served as the Lead Architect for Sensity Systems, Inc. before its acquisition by Verizon Communications in 2017.
Tackling Rough Patches and Preparing for a Better Future
According to Morgan Stanley, since 2010, manufacturers “have collected 2,000 petabytes of potentially valuable data but discarded 99% of it.” SWIM.AI was founded to solve for the challenge of building applications with real-time streaming data. Everything it has built over last 30 years assumes a batch analytics model, using a traditional database architecture. The database model still works, and works well, but it’s incomplete. As long as the organization is bound to batch analytics, everyone is doomed to throw away useful data. IoT requires a realtime edge computing solution to augment databases and provide a more comprehensive model which treats real-time edge data differently than historical cloud data. The 99% of data that’s being discarded today contains immense value. SWIM.AI acts on data locally to capture value from the 99%. Networks and centralized databases are still useful, but one can’t make them the bottleneck for acting on data and expect real-time results. SWIM EDX was created to solve this challenge, and provide a simple way to process, analyze, learn from, and react to real-time data at the edge.
Currently, SWIM.AI is building out a developer portal and supporting infrastructure (tools, SDKs, etc.) to make it possible for third-party developers to learn how to leverage SWIM EDX for their IoT and other real-time applications. SWIM.AI is already working with several channel partners to integrate SWIM EDX with their cloud and edge offerings, making the performance and cost benefits of SWIM EDX available to a wider audience.
Building Applications for Tomorrow
According to SWIM.AI nowadays, cloud-based batch analytics architectures dominate the IoT landscape. However, at the end of the day, it doesn’t matter how applications are being built for the present scenario. All that matters is what the applications of tomorrow will look like, and the use cases they will enable. Autonomous vehicles, AR/VR applications, gaming, artificial intelligence, Smart Cities, and many other “next-generation” technologies have yet to hit an inflection point, because they are stalled working with a paradigm unsuited for massively distributed systems which generate real-time data. Stateful, real-time edge computing will enable these applications, and more, by providing the means for processing and distributing that data throughout complex systems without slowing it down. The future of IoT applications won’t distinguish between the edge and the cloud. Instead, the future of IoT applications lies in acknowledging that there are different types of data, realtime and historical. Stateless, databasecentric architectures will always be optimal for performing batch analytics on historical data. But stateful, local computing is optimal for processing real-time data, and distributed architectures which operate state fully, at the edge, will come to define the next generation of IoT applications.