RESPUESTA INNATA: ¿UN TRATAMIENTO BASADO EN
EL INTESTINO HIPER-RESPONDEDOR DE LOS PACIENTES CELIACOS:
The component partition contains the object classes from which components are
instantiated. The enumeration put forward in this chapter represents the objects as static constructs and consequently does not give any indication when these objects interact. For RFID systems the concept of interaction between components is an integral process - interaction of tags and readers produces RFID data which is the basis for strategic level decision making. It is the RFID data which is used by the system owner, e.g. a company, to determine if certain events have taken place in the real world. However, before it is possible to think about what can be inferred from RFID data which has been produced when tags and readers have interacted, the next step is to consider the ways in which tag and reader components can interact.
This section models the associations between tags and readers in the association
partition. It models potential ways in which associations may emerge at the RFID data layer, between RFID components. As the data is generally assumed to be produced when a tag is at a reader, the associations which are modelled are representative of associations which have formed between a tag and reader, and at the layer below this, the real world layer, between entities which have tags and readers attached to them. As the data is being modelled in an RFID database, this section applies the principles of entity-relationship modelling first expounded by Chen (1976) to model these tag/reader associations to define a data view.
Figure 26 - RFID tag and reader associations
These have been modelled using ERD multiplicity concepts. A one-to-one (1:1) association is when one tag interacts with a single reader; other associations are built up from a 1:1 association when
additional tags or readers are introduced or when anti-collision is used at a reader. Thus, a hierarchical model is introduced to represent the ways that associations can form.
Figure 26 introduces the concept of a taxonomy of tag and reader associations which can be found in RFID data which characterise underlying tag and reader relationships. The rest of this section adapts the concept of multiplicity, in particular
the four standard multiplicity relationships (1:1, 1:M, M:1, and M:M), which exist in entity-relationship modelling, and shows that it is possible to model associations between tags and readers using the same multiplicity relationships. The knowledge imparted through the use of these terms is a controlled vocabulary to describe RFID structures – a key concept useful in allowing end users to identify domain abstractions (Arango 1994).
6.3.1
O
NE-
TO-O
NE(1:1)A
SSOCIATIONA one-to-one (1:1) association, illustrated in Figure 27, is instantiated in RFID data
when contact is made between one tag and one reader. For example, a tag has responded to a reader’s read command, in which case, data has been obtained from the tag. For the Mobilkom NFC payment system, which allows a mobile telephone to communicate with a single NFC-enabled cashier terminal to pay for goods (O'Connor 2005a), as the NFC-enabled telephone establishes a connection to only one cashier, and the cashier is capable of only one NFC telephone connection at a time, this is an example of a 1:1 association. Each time an association is formed, in this case, an RFID data record containing a timestamp and the component serial numbers is instantiated at the reader.
Figure 27 - One-to-one (1:1) association
The rest of the associations build on the concept of a 1:1 association between a tag and reader, but introduce more facets of the system which are available to the database if it was to look across multiple data records and data sources.
6.3.2
O
NE-
TO-M
ANY(1:M)A
SSOCIATIONA one-to-many association (1:M), illustrated in Figure 28, is instantiated in RFID
data when a tag has made contact with several readers in sequence in the system. The tag has been read, for example, at a reader and has then moved to another reader which in turn has also read the same tag. As these readers share a common database where they store their data, the database would be able to infer from the supplied data records, that the tag has been engaged with several readers over time.
An example of a 1:M association having been formed can be found in the Orlando/Orange County express way toll system which monitors vehicles via an RFID tag as vehicles travel along roadways which have readers mounted at certain locations (Swedberg 2004). As a vehicle uses a tag, which contains a unique serial number, over time this tag would have appeared at several readers in sequence, as the vehicle travels along the roadway.
Figure 28 - One-to-many (1:M) association
In this association it is important to recognise that at each reader a 1:1 association is formed, however, from the perspective of the database which has knowledge of all the readers, a 1:M association can be inferred. This is why the model shows associations in a hierarchical representation, as this representation depicts the concept that more complex associations are based on the basic 1:1 association.
6.3.3
M
ANY-
TO-O
NE(M:1)A
SSOCIATIONConversely, if a reader had several tags in front of it, and the reader is using an anti- collision protocol – these enable multiple tags to be read simultaneously (Finkenzeller 2004) - then it is possible that a many-to-one (M:1) association is be formed. This is illustrated in Figure 29. The anti-collision protocol schedules each tag in the reader’s vicinity to respond according to a scheduled time (Aloha-based protocols) or when addressed individually through their tag serial numbers (Tree
Walking based protocols). Instead of every tag responding at the same time period,
tags will respond according to how they have been scheduled, thereby avoiding collisions. But the reader could examine all records within a time window at a reader to determine which tags were active, and hence, constitute a M:1 association.
An example of this is the Dalsey Hillblom Lynn (DHL) Smart Box, which allows a physical entity to be identified via a tag when tagged entities are placed inside a container which contains a reader (Wessel 2007). There are many tagged entities and only one reader in this example and is thus a M:1 association.
Figure 29 - Many-to-one (M:1) association
As an anti-collision protocol allows many tags to communicate with a single reader at a time, conceptually, it is possible for a database to treat all tag data records as about a collection of tags, which in this section is termed a tag group. The formation of a tag group could be considered to be random or non-random. A random tag group is when an unrelated set of tagged entities arrives at a reader, such as when a group of vehicles congregate at a tollgate. Conversely, a non-random tag group is formed when products are arranged purposefully, such as when products are arranged on a pallet or in packaging. Recognising that tag groups are phenomenon reflected at the RFID layer through tag signals and also in RFID data, may allow for more effective recognition of groups in the real world.
6.3.4
M
ANY-
TO-M
ANY(M:M)A
SSOCIATIONA many-to-many association (M:M), illustrated in the full model in Figure 30, is
instantiated in RFID data when a tag is read across several readers which are using anti-collision protocols, and thus, could have read other tags at the same time as this tag. The M:M association subsumes all previous associations, and is thus, represented as the top-most association in the model.
An example of where a M:M association can be found is in the International Paper RFID system which monitors physical entities when they are placed on a forklift (O'Connor 2005b); as there could be many tagged entities on a forklift and many forklifts with readers in warehouses, this is an example of a M:M association.
A M:M association subsumes previous associations, while at the same time the structural properties of the system which have been built up can be decomposed to the most elementary 1:1 association. At the elementary level each tag is communicating with a reader through a 1:1 association. At the level of the M:1 association, anti-collision places each tag into a collection of tags read simultaneously. As a tag moves through a system, it could become a temporary member of tag groups at readers, or could already be in a fixed collection of tags. This movement through the system constitutes a 1:M association, while conformance to a collection of tags is a M:1 association. At various times in the system, different associations are formed, and different information can be gleaned from tags. The ability to glean such information is related to the system facilitating the establishment of these structures.
Finally, within the taxonomy a controlled vocabulary has also emerged as a way of describing the clusters of object classes that can be identified. Recall that Arango (1994) suggests that the derivation of a controlled vocabulary is beneficial to a domain model. It assists one to understand the domain at an object and data level. The concept of having 1:1, 1:M, M:1, or M:M groups of tags and readers is supported by a description of how they are formed through spatial and temporal enablers.
To summarise this section, what has essentially been described is a taxonomy of associations between tag and reader entities, enabled through the real world layer, but visible at the RFID layer in RFID data. The taxonomy abstracts classes of entities, like that in an OOA diagram, as at one end very specific object classes can be identified, to the other end where very general object classes can be identified. These are visible at the data level due to the hierarchy and information flow in RFID systems. These structures identify existing formations of tags and readers in systems - the advantage imparted here is the formalisation of these in a model. Knowing these structures exist, and what they contribute, assists users in realising their effects in systems.