Rugged Android handhelds and dead zones
A rugged Android handheld keeps working offline as long as the app on it can finish a task without a live connection. Zebra and other makers build these mobile computers for warehouses, yards, stores and field work. They run Android, carry a built-in barcode imager, and are managed as a fleet rather than as personal phones. They also work where coverage is patchy: Wi-Fi can end at the dock door and cellular drops behind concrete and steel, as How do warehouses and yards handle dead zones? describes.
Reading a barcode needs no network, because the identifier is printed in the label. The network is needed for the step after it, sending the scan to the warehouse or store system. So whether a handheld keeps working offline comes down to how the app receives scans, where it keeps them, how they move between devices, and how the app and its settings reach the fleet.
Barcode capture with DataWedge intent output
On Zebra Android devices, apps usually receive scans from DataWedge, a data capture service Zebra pre-installs on every Zebra Android device. DataWedge collects data from barcode scanners, imagers, magnetic stripe readers, RFID and other inputs. It works through profiles, each of which sets how it behaves for one or more associated apps, and profiles can be set by hand, through Android intent APIs, or deployed in bulk.
DataWedge has two common ways to hand over a scan. Keystroke output types the data into the focused field. Intent output sends it to the foreground app as the payload of an Android intent, so the app’s code can save the scan and queue it for upload. The app chooses an intent action and category, and the DataWedge profile is set to the same pair, which works like a channel the app listens on. The intent carries extras such as:
com.symbol.datawedge.data_string, the decoded datacom.symbol.datawedge.label_type, the symbology, such asLABEL-TYPE-EAN128com.symbol.datawedge.source, the input that produced it, which isscannerfor a camera, imager or scanner
Zebra offers four delivery methods: startActivity, startService, startForegroundService and a broadcast intent. The two service methods depend on the SDK version the app is compiled for. startService does not work for SDK version 26 or above, and startForegroundService does not work for 31 or above, so a current app receives scans in an activity or a broadcast receiver. The scan reaches the app whether or not Wi-Fi is up.
Store and forward on the handheld
A handheld keeps a scan through a dead zone by writing it to local storage before it tries to send it anywhere. Android’s offline-first guidance makes the local data source the app’s canonical source of truth, usually stored on disk in a database such as Room. For data that matters, it recommends lazy writes: write locally first, then queue the change to send to the network at the earliest opportunity.
That queue is an outbox. On Android, WorkManager is the API for work that must run reliably even if the user leaves the screen, the app exits or the device restarts, and Android gives periodic syncing with a server as an example. Handhelds are docked, rebooted and passed between shifts, so pending scans have to survive all of that.
Each record needs two more things to arrive intact:
- An operation ID. The server uses it to apply a retried upload only once.
- A capture time set on the device. The upload can arrive hours after the scan, so the record keeps the time the work happened. GS1 EPCIS separates the two with an
eventTime, when the event occurred, and arecordTime, when a repository recorded it.
Device-to-device sync and relay between handhelds
Handhelds can pass records to each other over short-range radio when neither one reaches a network. Android has built-in support for Bluetooth Low Energy, with APIs to discover devices, query services and transmit data, and its documentation names transferring small amounts of data between nearby devices as a common use. Wi-Fi Direct and Wi-Fi Aware connect Android devices over Wi-Fi with no access point.
Three patterns come up on a site:
- Handoff. One handheld passes a task, a pick list or a dock assignment to another, and the receiver confirms it has accepted the work.
- Shared state. Several handhelds count the same stock or update the same order, and their edits have to merge when the devices meet.
- Relay. A handheld in a dead zone passes its records to a neighbour, a vehicle terminal or a gateway closer to coverage, which carries them on. This is store and forward across devices.
The radio does not supply the rules. Discovery shows which devices are nearby, not which ones may receive a record. Android warns that data exchanged between devices paired over BLE is accessible to every app on the device, so an app carrying sensitive data needs its own app-layer security. When two handhelds change the same record apart, the app needs a merge rule or an authority, such as the warehouse management system, to decide. Android also limits what a BLE app can do in the background, which Can a BLE app run in the background on Android? covers.
Deploying apps with StageNow, MX and an EMM
Zebra fleets are staged with StageNow and configured through MX, often by way of OEMConfig and the company’s enterprise mobility management (EMM) system. An offline-capable app is an ordinary Android app, so it is installed and configured the same way as any other.
- StageNow runs on a Windows workstation. An administrator builds a profile with wizards such as Configure a Device, Connect Network, Enroll in an MDM and Manage Applications, then publishes it as staging barcodes or an NFC tag. An operator applies it by scanning the barcode or reading the tag with the StageNow client on the device.
- MX is Zebra’s XML-based framework for managing the capabilities and behaviour of Zebra Android devices. Devices running Android 13 still accept XML for staging, and the features added in MX 15.3 need Android 14 or later.
- OEMConfig configures Zebra devices from an EMM. It is installed from the Google Play Store together with Zebra’s schema, and works through Android Enterprise APIs where they exist and Android managed configurations where they do not.
- Managed configurations also let an IT admin set your own app’s options remotely. Android’s examples include whether an app may sync over cellular or only over Wi-Fi, the kind of site setting an offline-capable app needs.
OS updates and security patches reach Zebra devices through LifeGuard, generally once a month. DataWedge profiles support mass deployment, so the scanning setup an app depends on can ship with the rest of the fleet configuration.
Offline Protocol on rugged Android handhelds
Offline Protocol is device-to-device software for the layer above: handhelds sync records, send messages and hand off work directly, and relay for each other across a site, up to 8 hops over Bluetooth LE and Wi-Fi. Every message is end-to-end encrypted with MLS (RFC 9420) by default, and each device signs with its own key. Shared documents use edge sync, so edits made apart merge to the same result on every device when they meet.
The SDK supports Android 7 (API level 24) and later, through the React Native package or the Android library on Maven Central. Your app reads scans the way any Android app on a Zebra device does, for example through DataWedge intent output, and you deploy it with the StageNow, OEMConfig or EMM tools you already use.
Offline Protocol is a Zebra PartnerConnect ISV. The Zebra partner page covers scoping an evaluation on your own devices, and Logistics shows a handheld at the far fence completing a gate check through the trailers parked between it and the gate. The open-source demo apps include Stock Sync, a shared store inventory that keeps counting out of range, and YardGate, a gate check-in the officer approves on the spot.