Max Alexander, Co-founder at Ditto
Modern business apps are built to be cloud-dependent. This is great for accessing limitless compute and data storage capabilities but when connection to the cloud is poor or shuts down, business apps stop working, impacting revenue and service. If real-time data is needed for quick decision-making in fields like healthcare, a stalled app can potentially put people in life-threatening situations.
Organisations in sectors as diverse as airlines, fast food retail, and ecommerce that have deskless staff who need digital tools accessible on smartphones, tablets and other devices to do their jobs. But because of widespread connectivity issues and outages, these organisations are beginning to consider how to ensure these tools can operate reliably when the cloud is not accessible.
The short answer is that building applications with a local-first architecture can help to ensure that they remain functional when disconnected from the internet. But then, why are not all apps built this way? The simple answer is that building and deploying cloud-only applications is much easier as ready-made tools for developers help expedite a lot of the backend building process. The more complex answer is that a local-first architecture solves the issue of offline data accessibility but does not solve the critical issue of offline data synchronisation. Apps disconnected from the internet still have no way to share data across devices. That is where peer-to-peer data sync and mesh networking come into play.
Combining offline-first architecture with peer-to-peer data sync
In the real world, what does an application like this look like?
- Apps must prioritise local data sync. Rather than sending data to a remote server, applications must be able to write data using its local database in the first instance, and then listen for changes from other devices, and recombine them as needed. Apps should utilise local transports such as Bluetooth Low Energy (BLE) and Peer-to-Peer WiFi (P2P Wi-Fi) to communicate data changes in the event that the internet, local server, or the cloud is not available.
- Devices are capable of creating real-time mesh networks. Nearby devices should be able to discover, communicate, and maintain constant connections with devices in areas of limited or no connectivity.
- Seamlessly transition from online to offline (and vice versa). Combining local sync with mesh networking means that devices in the same mesh are constantly updating a local version of the database and opportunistically syncing those changes with the cloud when it is available.
- Partitioned between large peer and small peer mesh networks to not overwhelm smaller networks if they try to sync every piece of data. In order to do this, smaller networks will only sync the data that it requests, so developers have complete control over bandwidth usage and storage. This is vital when connectivity is erratic or critical data needs prioritising. Whereas, the larger networks sync as much data as they can, which is when there is full access to cloud-based systems.
- Ad-hoc to enable devices to join and leave the mesh when they need to. This also means that there can be no central server other devices are relying on.
- Compatible with all data at any time. All devices should account for incoming data with different schemas. In this way, if a device is offline and running an outdated app version, for example, it still must be able to read new data and sync.
Peer-to-peer sync and mesh networking in practice
Let us take a look at a point-of-sale application in the fast-paced environment of a quick-service restaurant. When an order is taken at a kiosk or counter, that data must travel hundreds of miles to a data centre to arrive at a device four metres away in the kitchen. This is an inefficient process and can slow down or even halt operations, especially if there is an internet outage or any issues with the cloud.
A major fast-food restaurant in the US has already modernised its point of sale system using this new architecture and created one that can move order data between store devices independently of an internet connection. As such, this system is much more resilient in the face of outages, ensuring employees can always deliver best-in-class service, regardless of internet connectivity.
The vast power of cloud-optional computing is showcased in healthcare situations in rural areas in developing countries. By using both peer-to-peer data sync and mesh networking, essential healthcare applications can share critical health information without the Internet or a connection to the cloud. This means that healthcare workers in disconnected environments can now quickly process information and share it with relevant colleagues, empowering faster reaction times that can save lives.
Although the shift from cloud-only to cloud-optional is subtle and will not be obvious to end users, it really is a fundamental paradigm shift. This move provides a number of business opportunities for increasing revenue and efficiencies and helps ensure sustained service for customers.
