RFID Middleware Design: Accomplishing real time integration

1
RFID Middleware Design Accomplishing real time
integration of an RFID Application with Oracle
Apps or any Business Application
2
AGENDA

• The concept of RFID
• Advantages offered and constraints imposed by
  RFID technology
• Key challenges in RFID Middleware Design
• Proposed design for an RFID middleware
• The Rule Engine
• Data filtering
• Data Aggregation
• Buffering and Exception Handling
• Data classification and dissemination
• Case Study Understanding the middleware design
  concepts through actual business transactions
• Summary

3
The concept of RFID

• RFID describes a class of technology that
  exchanges data wirelessly.
• RFID systems consist of three key components
• TAG, a microchip that contains a unique digital
  serial number and is attached to an antenna.
• READER, a device used to communicate with RFID
  tags to read data.
• SOFTWARE, that processes, routes and manages Tag
  data and Readers.

4

• The key characteristics that influence RFID
  performance are
• FREQUENCY determines RFID range, resistance to
  interference and other performance
• attributes. Most commercial RFID systems
  operate at either the UHF band (859 to 960 MHz),
• or High Frequency (HF) at 13.56 MHz.
• RANGE The proximity to the tag that a reader
  antenna must be within, to read the information
• stored on the tags chip varies from a few
  centimeters to tens of meters.
• SECURITY RFID chips are extremely difficult to
  counterfeit. One would need specialized
• knowledge of wireless engineering, encoding
  algorithms and encryption techniques to break-in.
• STANDARDS RFID standards exist for item
  management, logistics containers, fare cards,
• animal identification and many other uses. The
  International Standards Organization (ISO)
• and EPCglobal Inc. are two of the standards
  organizations most relevant for the supply chain.

5

• Advantages offered by the RFID technology
• Exceptional speed and accuracy More than a
  thousand reads can be performed each second.
• Versatility RFID enables monitoring and data
  collection even in extreme environments, without
• any manual intervention required.
• Tag-reader alignment Unlike bar codes, RFID
  does not require direct line of sight between tag
  
• and reader.
• Total Cost of Ownership The data on an RFID tag
  can be altered repeatedly.
• Constraints imposed by the characteristics of
  RFID
• Limited communication bandwidth.
• Reliability issues.
• Tag memory.

6

• Key challenges in RFID Middleware Design
• Seamless Data Bridging Maintain a channel of
  communication between the business application
  and the RFID application.
• Filtering Weeding out duplicates from the data
  flowing in.
• Data Classification Mapping the classified data
  to First the business application it should
  hit, and secondly the actual transaction
  bucket it falls into.
• Buffering and aggregation.
• Extendibility and adaptability.
• Interoperability To cater to the heterogeneous
  reader landscape.
• Flexibility Interfacing the data to the
  application in a format that is acceptable to
  that application.
• Two-Way data integration Writing back to the
  tags.

7
Proposed design for RFID Middleware
Oracle Apps
SAP
SCM
DATA PRESENTATION
RDBMS
RULES ENGINE
Filtering Data Classification
Buffering Aggregation
MESSAGING SERVER
Exception Handling
VTMS
PROTOCOLS
Protocol Engine
R3
R4
R1
R2
R5
RFID HARDWARE
Filtering
TAGS
TAGS
TAGS
TAGS
TAGS
8

• Rule Engine
• Well suited to the nature (large sets of data
  from multiple sources) and status (changing
  standards, protocols) of the RFID
  technology.
• Rule driven All forms of data handling and
  options to publish or subscribe the data are
  driven by the rules.
• Based on the concept of solving a problem using
  a set of logical rules specific to the problem
  domain.
• Converts data and messages from lower layers to
  actionable information for the upper layers
  based on the business or process semantics as
  perceived by the end user.
• Fits in perfectly for solutions requiring
  processing of large sets of rapidly changing data
  like in an RFID network.
• Offers immense flexibility to the users to
  incorporate their own rule chunks.

9

• Context Based Data filtering
• Filtering weeds our the inconsistencies in tag
  reading or writing like
• Multiple reads of the same tag.
• Some tags not being read.
• Erroneous reads.
• Removal of such tag read events based on the
  reader which generated the event and the tag
  data captured is the key objective of the
  filtering mechanism.
• Two filter types can be supported by the proposed
  middleware design
• Reader Identifier
• This filter type allows the application to
  specify that it is only interested data from a
  particular set of readers.
• Tag Identifier and Data
• The application can define the tag population
  that it is interested in, e.g., the restriction
  to
• tags attached to pallets.

10

• Data Aggregation
• - Reduces the flood of raw tag reads to more
  meaningful events.
• - Addresses the problem of temporary false
  negative reads and to smooth the data
  accordingly.
• Following aggregate types can be supported by the
  middleware design
• Entry Exit
• Reduces a number of successful reads of a tag to
  the best estimate when the tag appeared and
  disappeared from the read range.
• Count
• Applications can prefer to receive information
  about the total number of items of a specific
  category detected rather than the individual ID
  of each object.
• Passage
• When a tagged object passes a gate, applications
  would prefer receiving a passage event rather
  than being forced to interpret a sequence of
  entry and exit events from two individual
  readers.
• Virtual readers
• When an application does not distinguish between
  two readers, this aggregate type allows it to
  virtually join their read range.

11

• Buffering
• Data persistence component of the middleware,
  which is based on message queues.
• Buffering is required for two main purposes
• Facilitates asynchronous processing of the data
  streams. This is imperative for providing
  sufficient time for the rules to add value to
  the raw data.
• Supports different latencies of the destination
  applications, depending on the application type.
• Exception Handling
• Raising the red flags Any discrepancy of data
  during the data scrubbing is processed as
  exceptions. Numerous alerting systems are
  available for resolution emails, messages, or
  user defined triggers.

12

• RFID data dissemination
• - Classifies the extracted data into relevant
  transaction types.
• - Classifies the transactions to the relevant
  business application.
• Full content-based routing
• Key element for classifying the data into
  correct transaction buckets.
• Full content-based routing (rather than subject-
  or topic-based routing) is accomplished in order
  to carry out the filtering within the messaging
  system itself.
• Air Interface Subscription feedback mechanism
• Overcomes the constraint of limited bandwidth
  available to RFID.
• Feeds back the readers to communicate whether
  applications are interested in
• the RFID data they produce.
• Leads to appropriate adaptation of the queries
  exercised by a reader over the air interface,
  (e.g. targeting a particular tag population at a
  higher sampling rate or switching off completely
  to make the bandwidth available to another
  reader).
• The filtering of the RFID data is then no longer
  carried out in software, but over the air
  interface.

13

• VTMS Reading from and writing to a tag
• VTMS is specifically aimed at shielding the
  application from the particularities of RFID tag
  memory
• limited memory size.
• different memory organizations.
• reduced write range.
• VTMS response to different write scenarios
• If the write succeeds, the RFID middleware will
  acknowledge this to the application and will
  store a
• backup copy of the data in the virtual
  representation of the tag in the VTMS.
• If the memory gets corrupted at a later stage or
  the application wants to access the tags memory,
  
• while the tag is outside the range of any
  reader, the RFID middleware can make the data
  available
• via this virtual memory.
• If the write to the tag fails due to
  insufficient power, the key-value pair will be
  stored in
• the VTMS and flagged as open. The RFID
  middleware will retry the write command at a
  later
• point of time.

14
Case Study Deployment of RFID in the Supply
Chain
Manufacturing WIP Status SUPPLIERS MFG - WIP
Pick release and RFID tag attachment SUPPLIERS
Ready to Ship
Finished Goods SUPPLIERS MFG - Complete

• Installed at site
• Scan/Verify
• Installation, Periodic meter data
• for maintenance
• Asset Management

Transport enroute to stocking location Shipment
Intransit

• Port of Import (POI)
• Scan/Verify
• Unload On Hand at POI
• Departure from POI
• POI- In/POI- Out

• Receiving at Receiving dock
• Inspect all components installed
• Sales order shipment
• Warehouse

In- House Mfg. WIP Status SHOP FLOOR MFG - WIP
Transport enroute to Customer site Shipment
Intransit
15
PO receiving against ASN
ASN
Supplier

• Shipment Number Vendor Name Airway Bill Number
• Transaction Date Quantity Unit of Measure
• Item Description Document and Line Number Vendor
  Site
• Bill of Lading Packing Slip Number of
  Containers
• Additional data generated by the reader at the
  Receiving dock
• Location Code (Row / Rack / Bin)
• Reader Id
• Transaction Date
• The filtered transaction then can be classified
  as a PO receiving transaction. This can be done
  by a combination of the tag id, reader id,
  location at the receiving dock and PO number.
• TAG ID READER ID PO NUMBER LOCATION AT
  RCV DOCK

PO RECEIVING TRANSACTION
16

• Intra warehouse movements
• Movement from one stocking location to another
  within the warehouse.
• Movement from the warehouse to a manufacturing
  facility.
• Movement to a packing and staging area for
  shipment against a sales order.

Manufacturing Facility
Transaction Type Transaction Date Bill of
Lading Physical Location Packing Slip Demand
Source Shipment number Quantity and UOM Tag
ID Warehouse Location Code Transfer
Subinventory Sales Order Number LPN Delivery
Number Transfer organization Combination of
these fields along with the tag ID and reader ID
identifies and classifies the type of business
transaction. TAG ID READER ID SALES
ORDER LPN DELIVERY NUMBER
SALES ORDER BASED SHIPMENT
17
Application based filtering and
classification The transaction classification
along with the RFID data will be used to identify
the business application bucket in which it
resides. READER LOCATION
TRANSACTION CLASSIFICATION
DESTINATION APPLICATION Receiving Dock
PO Receipt ERP/WMS Warehouse
Intra warehouse movement
ERP/WMS Supplier Site Customer
Service CRM Once this classification
is completed, the data will be mapped as per the
requirements of the destination application.
18

• Summary
• RFID is a promising technology that could yield
  immense contextual intelligence.
• RFID middleware forms the bridge between the
  RFID Application and the pre-existing business
  applications.
• RFID can be deployed at each stage of the Supply
  Chain for improved visibility of goods.
• The proposed RFID middleware design
• - Talks to the RFID application and extracts
  data.
• - Performs context-based Data Filtering.
• - Classifies the extracted data into relevant
  transaction types.
• - Classifies the transactions to the relevant
  business application.
• - Performs Buffering, Aggregation and Exception
  handling.
• - Adheres to the relevant protocols.
• - Is Interoperable and Scalable.
• - Overcomes the limitations of Tag Memory
  through VTMS.