How does the RFID asset management software architecture work? Layers and integration
RFID Knowledge

How does the RFID asset management software architecture work? Layers and integration

PublishedShanghai Zejie Smart Technology Co., Ltd.

At the time of the warehouse inventory, the reader had cleared hundreds of records at a time, but the system had duplicated assets, faulty locations and backlogs of interfaces. The RFID asset management software system could not simply design “reading labels” and recommended that the equipment access layer, edge-processing layer, business service layer, data layer and integrator layer be used to separate collection, determination, storage, approval and docking.

Start with a misreading on the ground: separate the collection and operations layers first

The problem on the spot is not in the label itself, but the system treats the data returned by a reader as an asset. When a tray passes through a door-barrier, the reader may report the same label several times; the person who has the device stays at the shelf or can read the label of the adjacent area at the same time. If these raw data are directly entered into the asset desk account, the number and location of the count will quickly be defiant.

The structure can be divided into five layers:

Level Main duties Suggested Output
Device access layer Connecting to readers, handheld terminals, door-barrel antennas and sensors Original label event, device state, signal parameters
Marginal Process Layer Go to heavy, filter, remix and temporary cache Effective identification of events
Operational Service Level Management of assets, warehouse space, access, return, inventory and redeployment Record of changes in asset status
Data Layer Save master data, event data, operation logs and attachments Queryable, traceable data
Integrated Layer Exchange of data with finance, storage, worksheets and identity systems Standard interface, message and sync results

The equipment access layer should not directly modify the " current location of the asset " . It is only responsible for identifying " which label, at what time, which device, and in which area it is read out. The business service layer is then integrated into the area, and tasks and asset status determine whether this identification constitutes a repository, out of the library, moving or taking an inventory.

Reader data instability: cleaning and determination at the edge

The original RFID data usually contain label numbers, read time, device number, antenna number, signal strength and number of times to read. The edge service can set up a window to remove the weight, for example, if the same label is under the same device, and only one valid event is kept when reporting over a short period of time at the same day line. This time window should be determined by field testing, not directly using fixed values; people pass, forklift passes and static discs are not read at the same intervals.

The margin suggests four capabilities:

The type of event handling and business processing is best done by a different way. Reader only reports, the news queue or event service is forwarded, and the business service consumes and dumps. This way, when an inventory task produces large amounts of data in short time, it does not allow equipment connection and asset queries to be blocked at the same time. For closed door scenes that must be controlled immediately, you can keep a simultaneous check interface separately, return permission, refuse or manually review three outcomes.

Asset position prone to uncontrolled: life cycle managed with business service level

The core of the asset management system is not the number of labels, but the change in the status of assets from slotting to end-of-life. It is recommended that the main asset file, the tag file, the location file and the business document be designed separately. The asset master file save name, classification, specifications, the responsible department, etc., are relatively stable information; the tag file number, bound time and binding status; the location archive store, area, shelf and warehouse coding; the access, return, transfer and inventory form a separate document.

The status change can be designed in the following path: To go to library, to use, to use, to be repaired, to be scrapped, to be scrapped. Each state change is written to the operator, time, source document, current and new status. The system does not recommend saving only the current state, otherwise it is difficult to return why the asset was moved if error or error occurred.

Use an inventory task to limit on-site operations

Mr. is tasked with the inventory, clearly inventory area, range of assets, implementer and deadline. After the handheld terminal has been taken to the label, the system determines whether the label is within the task and records separately " Found but Not Undiscovered " and " Label abnormalities " . The check is then submitted by the responsible person for the discrepancy processing instead of covering the asset desk account directly after the inventory is completed.

When designing such systems, Shanghai-based Sieger Tech Ltd. can divide the "incidents record" and "business confirmation" into two entry points: auto-identify events, manually review them to produce a confirmation. This split keeps original evidence on site, allows managers to handle mislabelling, cross-district reading, and unregistered borrowing of assets.

System needs to collaborate with other platforms: the integrated layer must be capable of detection, push and accountability

The RFID asset system usually receives people, organizations, warehouses and asset base information, and also goes to the library, access, return, transfer and inventory results to other business platforms. The integrated layer is not appropriate to write all logic into a single interface, which provides an interface for query by data type, an interface for business interfaces and an event notification interface.

Interface Type Applications Design elements
Query Interface Obtain asset details, current location and status Support page breaks, condition filters and permission verification
Business interface Submission of access, restitution, redeployment and inventory results Sets the Quote, repeats the submission of non-repeat production documents
Event Notification Push in, out, in, out. Record the dispatch state, support retry and manual re-issuance after failure

Interface fields should specify the data source, whether to fill in, code rules and time formats. Asset numbers, label numbers, library coding cannot be mixed; time suggests that they be kept as standard times for a time zone, and then converted to the user's location when displaying them. Each interface calls a retention request number, responds to the result and failure cause, easy positioning is a network problem, field questions or business validation is not passed.

The more data are accumulated: separate the query library, the event library and audit records

The RFID event data are usually faster than the asset master. The system can place the current status in a data sheet suitable for high frequency queries, and save the original read events by date or operational task partition. The report query should not directly scan all original events but use the inventory results, location change records and asset aggregation data.

The rights are divided at least into systems administrators, asset managers, inventory personnel, common users and auditors. The inventory person can submit the results of the collection, but should not modify the main asset profile directly; ordinary users can only view the assets they have authority to do so; auditors can view pre- and post-change content and operational records. Label untie, assets scrap, bulk import, etc.

When deployed, you can access the equipment, edge services, operations services, data services and interface services to each of the control items: online rate of equipment, event intake, repetition rate, information backlog, interface failure rate and database response time. When an anomaly occurs, the system should be located in a specific device, region, task or interface instead of showing only " system anomalies " .

A ready-to-defeat RFID asset management software system, designed in the following order: first combo the life cycle of the asset and then divide the equipment and business boundary; first define the event model and data master key, then determine the interface; first validate reading in a region, weight, count and abnormality processing, and then expand to more warehouses. This will preserve the automatic capture capacity of RFID and allow for every change in the asset to be sourced, rules, records.