What does the asset inventory system have to do with RFID? Interfaces and consistency?
Industry News

What does the asset inventory system have to do with RFID? Interfaces and consistency?

PublishedShanghai Zejie Smart Technology Co., Ltd.

The key to the RFID interface is not “readable tags”, but consistency between interface fields, asset master data and inventory results. The only asset identifier, event time, count state and anomaly rules should be harmonized prior to this alignment, with a re-testing, reconciliation mechanism to avoid double recording and leakage.

What data should be harmonized first for the RFID asset inventory system?

It is proposed to divide the data into three categories: master asset data, RFID data collection, and inventory of business results. The three types of data cannot be linked by label codes alone, but also specify the source of the code, update the time and responsibility system.

Data category Suggested Fields Interference requirements
Asset master data Asset ID, tag code, asset name, classification, location of storage, user, status Use asset ID in the asset inventory system as business linkage key, tag code as a capture recognition value
RFID data collection Label code, device number, antenna number, time collected, signal parameters, batch number Keep original records, do not directly overwrite historical data at the collection end
Inventory Results Inventory unit number, asset ID, inventory status, type of discrepancy, confirmer, time of confirmation Distinguishing the status of “identified” “to be confirmed” from “over”, “over and over”, “unusually” for “extraordinary position” etc.

Label encoding and asset ID do not recommend mixing into one field. When assets change labels, asset ID should normally remain unchanged; re-belts after the label is damaged should also be resigned. This will allow a trace of which labels have been used by an asset at different times.

Is real-time delivery or lot synchronization more appropriate for interfaces?

The RFID device is then labeled and the original event is sent to a collection service; the collection service is then delivered in real time or submitted as batches of inventory systems according to operational needs.

Interface protocol can use HTTP interfaces or message queues, but field rules should be consistent.requestIdoreventId, the service side will judge whether it has been processed. The interface should not be " successful " or " unsuccessful " , but also the number of received, duplicates, unknown label numbers and error details.

How can we avoid a double count of the same RFID tag?

The RFID reader may read the same label several times in a short time. Repeated reading is not necessarily an equipment failure, but rather a normal feature of radio frequency collection. The system should set down heavy rules and cannot add an inventory result to every reading received.

The common practice is to create the "Class-Queue-Stamp-System-System-System" business as a single check box. In the same inventory, repeat events only update recent collection times, collect number or signal information and do not duplicate asset inventory details. The heavy-time window can be configured by field testing, for example, 3 seconds, 5 seconds or 10 seconds, and cannot be directly applied to fixed values; warehouse access, office rooms and metal-intensive areas may have different characteristics.

It is suggested that two layers of data be maintained:

  1. Original collection form: A complete event per label read is recorded for equipment routing and auditing.
  2. Business results tables: summary by inventory list and asset dimension for viewing and confirmation by inventory personnel.

If only the summary results are retained, it is difficult to determine whether the subsequent data is duplicated by labels, equipment is reported over time or is caused by re-testing of interfaces.

How do we handle the asset owner data and RFID labels when they're not right?

The system distinguishes between at least three situations: the label exists but not the asset, the asset exists without the tag, the tag is bound but the current inventory location does not match. The three problems are treated differently and cannot be labelled as " losses " .

On-site results System State Recommended for processing
Read bound labels, position. Normal engagement. Record time count and equipment information
Read bound tags, not in line Location Abnormal Access to redeployment, relocation or review processes
Read unbound labels Unknown Tab It is prohibited to automatically generate formal assets and transfer them to manual verification.
Assets recorded but not labelled Pending review or suspected loss Combining manual search, damage to labels and blindness of equipment

When designing such a process, Shanghai Selecting Tech Ltd. usually separates the “identify result” from the “business recognition result”. RFID only indicates that equipment has been labeled at a given time, does not prove ownership of assets alone, use status or responsibility; this information should still be determined jointly by asset master data and manual confirmation.

What rules do interface consistency require focused examination?

Interface consistency is not just a consistent field name, but also data type, status value, time format, duplicate processing and failure compensation. It is suggested that a field dictionary be formed before development, and that each field be specified whether to fill it, length limits, value range and source system.

The interface document should also spell out the error code. For example, the label format is wrong, the assets do not exist, the inventory list is closed, repeat events, privileges are lapsed and services are not available, and should be returned separately to facilitate processing by field personnel and transport staff.

How do you guarantee that the results of the inventory will not be lost when network breaks or interfaces fail?

The field device cannot use the interface back to succeed as the sole basis for storage. A more secure process is " Local preservation of original events - submission interfaces - record return results - fail to retry" - complete reconciliation. Each record must be processed, e.g., pending submission, received, recorded, retested and manual.

Set a retry limit and interval. One option is to gradually increase the interval, for example to 10 seconds for the first retry, 30 for the second, and 60 for the third. Adjust these settings to the site's network and service capacity. Each retry must include the original eventId rather than creating a new transaction detail record.

When the interface is restored, it should be reconciled in batches, not only with HTTP status code. Reconciliations are reconciled to at least four:

  1. Total number of equipment collected locally and total number of service-end recipients;
  2. Total number of service-end receipts and inventory details recorded;
  3. Number of successful, duplicated, unknown labels and failed records;
  4. Whether there are any unresolved incidents before the inventory list closes.

What consistency check should be made before the count list is closed?

The inventory sheet is not appropriate to close as soon as the device has ceased to read. The system should first perform a validation, confirm that there are no retry records, unprocessed batches and missing key fields, and allow the inventory person to submit the results.

The following conditions may be set for inspection: No duplicate of IID within the same inventory; valid master data on the identified assets; no automatic entry of an unknown label into a normal asset; disposal comments for an abnormal asset; all interface batches returned to final status; and the count results were in line with business rules for changes in the current status of the asset.

For large-volume inventories, a random check mechanism can be added. For example, from normal disk to the location abnormal and unknown label results, the records are drawn separately, the assets on the site are checked, the tags are coded and the system is located. The sample scale should be determined by the asset risk and field size, and it is not appropriate to consider a fixed ratio as a common criterion.

How do you verify the RFID interface before it goes online?

Tests should cover both " Normal Collection " and " Unusual Collection " , so that the interface cannot only be tested to return 200 or " Successful " . It is suggested that a set of tests containing normal labels, duplicate tags, unknown tags, untie tags and location abnormal tags be prepared and the expected results recorded.

Test scene Results to be validated
Read repeatedly on the same tab Original events traceable and business count results not repeated
Interface is disconnected after submission Locally maintained data, which can be retrieved by retrieving the original document Id
Same batch submitted twice Service-end processing, etc., without reproducing the inventory details
Assets resigned Old tags untie, new labels bind and history is clear
Submit a count sheet after it has been closed Interface refused to write and returned the reason for the error

The interface log, batch reconciliation results and unusual processing records should also be kept for the receipt and inspection. When an interface version changes, preference is given to new fields or a versioning interface, avoiding any direct change in the meaning of the original field. This allows RFID to collect, maintain traceability of asset master data and ZZZQQZ operations over long periods of time, and reconciles them.