weeve Documentation

The edge explained

The word 'edge' is a crucial concept in understanding the weeve platform, but may lead to confusion - how 'edgy' is your application?

Where is the edge?

In the context of edge compute, the word 'edge' is ambiguous. Edge computing is computing that is done at or near the source of the data. Edge compute can be contrasted with cloud compute. Given this definition, it's clear that there is ambiguity in the term edge!
It's helpful to consider the journey that data takes in a generalized end-end IoT solution. A generalized view of the data journey is presented in figure 1 below, from generation or consumption at the end devices or sensors, all the way to cloud processing and data presentation. Along the journey, the data is routed through a series of hardware devices. In some contexts, industries, an solutions, each of these device categories can be considered 'the edge'. The 6 locations along the journey are described below, with a perspective on how the weeve platform would support at each location, the first 5 locations being a potential 'edge', and the 6th location being the remaining data processing and storage in the cloud abstracted as 'core'. Of course, this location itself contains a potentially complex data processing pipeline.
Figure 1 - Edge compute definition (Modified from BCG - The Battle at Computing's Edge, March 15th 2021)
  1. 1.
    Device edge - Onboard signal and data processing. This edge typically handles only a specific application purpose. Because of this limitation in purpose, this edge is often, but not always, characterized by constrained computing resources. Therefore, this edge location may not be suitable or able to by weeve as it is often too constrained to run edge applications.
  2. 2.
    Gateway edge - The next immediate step in the data journey, is a point of aggregation connecting many devices with the internet. The device may transform data from a machine specific interface and protocol, into internet protocol (Gateway function). They may act as a pure TCP/IP router, routing traffic between a LAN into the WAN (Router function). 'Smart' gateways are gateways that are less constrained, and may run an application operating system. The weeve IoT platform is well suited to operating on smart gateways.
  3. 3.
    Premise edge - Some companies own a group of servers that are privately owned and operated. These 'on-prem' servers are sometimes configured to operate as a "compute cluster" using container orchestration. Some on-prem solutions are architected to operate in a hybrid mode, incorporating cloud solutions.
  4. 4.
    Access edge - These devices are designed for high-speed routing and may not support generalized applications. In some instances, a parallel compute device is co-located with the access edge hardware.
  5. 5.
    Metro edge - In internet topology, a regional routing of traffic occurs at a Point of Presence (PoP). The PoP is a collection of telecommunications technologies and equipment that allows users to access the Internet. They are typically located within data centers near large Internet Service Providers (ISPs). For weeve, use cases are currently not developed as this network compute infrastructure is mission critical, regulated, and built for a limited, non-general purpose.
  6. 6.
    Core - Since weeve is built around container orchestration, the weeve platform can seamlessly be operated in the cloud on dedicated instances or in a cluster setting.
This simplified graphic can be used as a starting point for clarifying discussions around 'edge compute' and where the weeve platform would be most advantageous. In many solution cases, not all edge locations are present. For example, in telecommunications, end-devices often communicate directly with the Access edge (a cellular network tower). In such a case as bike sharing micro-mobility over a 3G network, there is no Gateway or On-premise edge. In other cases, data may be encrypted at the source, and limit the ability to do any data transformation at the edge.
Any discussion regarding edge compute and how weeve brings value to end-end solutions is enriched by first understanding the generation and flow of data through the edge and into the core. Since the weeve platform is built on generalized technology, it is possible to integrate the weeve edge compute platform in any of the above described locations.