Smart Mapping: Optimizing Compute Across Cloud and Edge in IoT Networks

Mapping computational tasks across diverse IoT networks has traditionally been manual, leading to architectural lock-in. This talk explores an innovative approach to developing distributed applications that integrate cloud infrastructure, edge devices, and IoT nodes. We’ll discuss mapping compute tasks as Erlang Processes within a dataflow graph onto a heterogeneous network of nodes, guided by criteria such as:

• Statically fixed node mappings
• First-bit latency
• Network and compute bandwidth
• Latency influenced by bandwidth limits
• Energy consumption
• Specialized compute needs
• Data sensitivity and trust We’ll demonstrate this through live demos with cloud and IoT nodes, highlighting how these mappings make use of a compute continuum, meeting all constraints and adapting to changes.

Join us to explore the intricacies of compute mapping in heterogeneous IoT networks and learn techniques for developing advanced distributed applications leveraging both cloud and edge resources.