RFID Software Explained Part 3: Implementation Architectures
In Parts 1 and 2 we described RFID Device Management Software and RFID Business Applications. Today, the topic shifts to implementation architecture. This is a set of decisions about how you deploy RFID software and where it resides. It also can have a big impact on your server infrastructure cost and placement as well as overall system scalability.
Connecting the Systems and getting RFID Software off the Ground
RFID software and large existing systems can both present challenges to initially deploying RFID and scaling the production solution . The key to both short and long term success is breaking down the problem into logical pieces of functionality. To do this, there are a few key questions:
1. Exactly what data do I need from the reader?
2. Is my current business system capable of handling serialized data?
3. Are there workflow changes that RFID creates that I cannot manage with existing systems?
As you answer these questions, you will want to think about challenges in both short and long term, since they will be different.
Direct to the Back-End RFID Integration
Let’s dive into two examples: In the first example, we consider a distribution facility using a modern Warehouse Management System (WMS) which currently handles serialized data. In this situation, RFID may be used at dock doors with RFID portals and on conveyors with portals of RFID tunnels. We may want to see data on tagged items (perhaps IT equipment, medical devices or other inventory) and ensure that the materials are being routed through the facility properly. After the items are consolidated for shipment, we then want to ensure that it is being shipped to the correct location.
From a software perspective, it will be important to instruct the RFID reader how to separate items or transactions on the conveyors. It’s likely that a PLC or photoeye will trigger the read start and the reader will need to either read until another photo eye is tripped or for a specified period of time. Filtering of the data will need to uniquely identify the item and then package it (RFID serial information) up and send it to the existing system.
At the dock door, there will need to be a way in the existing system to open up a shipment and then the RFID system will need to package together a batch of reads going through the door for comparison on against a list in the WMS for the shipment.
Because the existing system is advanced and already has serialized data processing capabilities as well as routing and shipment comparison capabilities, no special RFID-specific screens will be required. The RFID edge software can communicate directly to the WMS.
Using an agent based software like EasyEdge™ allows you to add RFID read points at will and scale to additional locations quickly and easily. There will be minimal incremental load on the network system because the filtering takes place at the readers and there will be no need to have distributed servers since the agent is loaded directly on the reader.
Integrating to a Legacy System when RFID Business Application features are needed
Although it would be ideal for the RFID readers to always connect directly to the installed enterprise back end systems, often these systems are not capable of supporting the serialized data that RFID provides. In addition, there are typically new workflows introduced with RFID that require new user screens or functionality. In these instances, it is faster to get a project off the ground by implementing some lightweight RFID-specific business application software. If desirable you can then transition the functionality over time to a direct connection as the existing systems catch up.
In our second example, we will consider a system that manages items at the batch level, but is incapable of managing at the item level. The RFID system will again need to read items on conveyors and out the dock door, but new functionality will be required to map serialized item and batch information. In addition, some screens will be required to handle exceptions.
In this case, an agent based RFID software model works again at the edge, but a new piece of functionality will be required. If that function is created using JEE or .NET in a standard application server, it will be possible to phase out this piece of functionality over time and move toward an architecture similar to the first example.
RFID System Scalability
By utilizing the above architectures, enterprises have the flexibility to scale their RFID systems in a way that is consistent with standards based enterprise architecture. End users also have the ability to treat RFID as it should be - as an alternative data collection architecture rather than an alternative enterprise architecture. The key to RFID success is implementing the entire system in a way that can start reasonably small and scale appropriately and consistently over time.