Then, using a stateful architecture we can allow the streaming data to build the model of the real-world, on-the-fly, as the data is received. A stateful digital twin model of the edge can be efficiently managed in memory, and updated at the rate at which data is received (whereas cloud-hosted big-data stores are slow). In this architecture – also known as the distributed actor model – digital twins represent entities in the data, and statefully process data in real-time from their real-world siblings – devices, applications and other infrastructure. As data arrives, each digital twin processes its own data – cleaning and analyzing it..With this architecture, all that a user needs to do is specify how each twin should analyze (and even display) its data. Digital twins can perform simple analytics, or collectively drive inputs into much more complex analytical functions, including self-training and inference using complex DNNs..Digital twins can also collaborate to perform sophisticated analyses that require inputs from multiple entities. These could include simple “joins” of the evolving state of different twins, sophisticated correlations and even machine learning.. The specification of what is required needs to be easy for a traditional developer, with skills in a high level language like Java, who can simply invoke functions for analysis or learning from any major toolset..And what of the original data?.Well, if you need it, save it. But in the vast majority of use-cases for streaming data, there is no need to save data that is only ephemerally useful..Instead, digital twins can learn on the fly, at the edge..Succeeding with AI in today’s enterprise requires us to free users from the challenges of infrastructure needed to collect, store, analyze and model data. It requires an abstraction at the level of digital twins and how they relate and share context – to allow users to easily describe their environments and the relationships between entities in order to derive insights. SWIM provides that abstraction and takes care of all the complexity underneath: Building stateful digital twins that learn from the streamed data of their real-world siblings and describing analytics, training and inference that are required based on the observations of multiple twins that build a smart infrastructure..About the Author Simon Crosby is CTO of SWIM.AI..Simon brings an established record of technology industry success, most recently as co-founder and CTO of Bromium, a security technology company..At Bromium, he built a highly secure virtualized system to protect applications..Prior to Bromium, Simon was the co-founder and CTO of XenSource before its acquisition by Citrix, and later served as the CTO of the Virtualization and Management Division at Citrix.. More details