Fig 2 - uploaded by Dmitry Zaitsev
Content may be subject to copyright.
The state diagram of piconet devices [2] 

The state diagram of piconet devices [2] 

Context in source publication

Context 1
... the initial stage of development, due to high cost of devices, the Bluetooth [ 1, 2] protocol was mainly involved in military and special applications as the basic protocol for mobile sensor networks. It was used to collect the information obtained by autonomously working sensors at a radius of about a kilometer for visual observation, audition, radiological monitoring etc. In [3] the application of the Bluetooth protocol for battlefield panoramas creation was also mentioned. Recently, as a result of reduction of prices for Bluetooth devices, the protocol finds wide office applications for construction of wireless piconets. Moreover, we now have a wide choice of computer peripheral devices supporting information exchange according to the Bluetooth protocol like: printer, keyboard, mouse, acoustic system etc. Also these days hand free mobile phones have become part of one’s basic need. As the construction of analytical models for the Bluetooth protocol is quite difficult due to high complexity of technology, the simulation technique is a prospective direction for investigation. There are well known applications for the specialized simulation systems like NS for the investigation of Bluetooth energy-saving modes [3]. But the application of specialized simulation systems possesses a series of disadvantages; many of them are in the models integration in heterogeneous networks. Colored Petri nets [4, 5] of simulation system CPN Tools [6] are the universal algorithmic system which allow the modeling of telecommunication devices and networks [7, 8]. Earlier proposed method of measuring fragments [8] provides the measurement of nontrivial characteristics behavior on simulation. The goal of the present work is the construction of typical models of Bluetooth master and slave devices in the form of colored Petri net and to estimate the effectiveness of protocol address space usage. Recently the selling of Bluetooth equipment considerably exceeds that of IEEE 802.11 equipment – well-known standard of Wireless Local Area Networks (WLAN). Therefore, IEEE organization has created a special group for the development of 802.15 standard mainly based on Bluetooth specifications named by standard of Wireless Personal Area Networks (WPAN) (also called a Piconet). Special Interest Group (SIG) founded in May, 1998 by companies such as Intel, 3COM, Ericsson, IBM, Motorola, Nokia, Toshiba are also involved in the development and maintenance of Bluetooth standards. The basic specifications of the protocol are presented in documents [1]. The architecture of the protocol is represented in Fig. 1. The interface of the protocol with the air is presented by the Radio Frequency (RF) layer. The nominal power of the antenna is in the range of 1-100 mW, which provides 10-100 m radius of operation. The protocol uses frequency hopping in the unlicensed range from 2.402 to 2.480 GHz. There are 79 frequency levels (channels) in the mentioned frequency range. The Baseband layer provides a random sequence of channels for transmission of information and a special procedure for sequence concordance. The random sequence of channel hopping guarantees a joint operation of few piconets in the same frequency range. Each piconet is constituted by one Master device and a few Slave devices. The packet are transmitted in slots (1, 3 or 5) where, transmission of each slot requires 625 μs (each of which may be transmitted through its own channel). The standard frequency of channel hopping is 1600 channels per second, which provides the rate of information exchange to 721 Kb/s. Each device has unique 48 bits address compatible with IEEE 802.15. For coordinated pattern of frequency hopping the Global identifier (Global ID) of piconet is used, which the master device communicates to all the slave devices. The state diagram of piconet devices provided by Link Management Protocol (LMP) is represented in Fig 2. The device not attached to piconet is turned to be in Standby state. In this state the device, which is listening can announce the existed piconet (Inquiry) or request an attachment to piconet (Page). For attachment to piconet master device sends the packet Page containing Global ID. After the attachment to piconet, 3 bits Active Member Address (AMA) is allocated to the device, which is used for data transmission. Thus, there are no more than eight simultaneously active devices in the same piconet. Due to necessity of information exchange with new devices, master device sends the packet Park to one of the active slave devices compelling it to return its AMA into pool and assigns Passive Member Address (PMA) to it. Thus, there are no more than 256 attached devices in a piconet. There are three energy-saving states: Park, Sniff, and Hold. In the Hold state, the device does not release its AMA. However, in the Sniff state, the device may transmit data at certain intervals of time. Bluetooth controllers have to support the standard Host Controller Interface. In the Connected state according to Logical Link Control and Adaptation Protocol (L2CAP) two kinds of data packets are provided: Synchronous Connection Oriented (SCO) and Asynchronous Connectionless (ACL). SCO packets requires connection establishment and are mainly used for the transmission of isochronous data for instance audio. On the other hand, ACL packets are used for transmission of brief messages. For rendezvous with the protocols of application level a series of auxiliary protocols are used for instance Service Discovery Protocol (SDP) protocol for serial cable emulation RFCOMM etc. CPN (Colored Petri net) Tools [6] embodies the simulation of Petri net graph [4] and programming language CPM ML [5] for attributes description of net elements. Petri nets, in which a token is not elementary but may possess individual characteristics, are traditionally named by colored nets. In CPN Tools a token is an object of abstract data type. Moreover, attributes of transitions are specified to check or transform the characteristics of tokens. Model of Bluetooth slave device is represented in Fig. 3. An essential item which comprehends model organization is the description of used types, variables and functions as represented in Fig. 4. In the present work, the main type (colset) of tokens is pkt (modeling packets of data transmitting among master and slave devices). Notice that, only headers of packets are represented in the model, which contain AMA of sender srcama , AMA of recipient dstama , type of message mtype and PMA of slave device pma for only some types of messages. Addresses and the type of following type of packet are modeled by integer numbers: allocation of AMA (mtype=1), release of AMA (mtype=2) and data transmission (mtype=3). Type RF models the data transmission in the air and provides the control of transmission order in such a way that the air may be available for transmission by master device availmaster , transmission by slave device availslave or busy with transmission of packet r . Let us consider the organization of the model. Contact place myPMA holds PMA (and myAMA holds AMA) of slave device. The initial marking of myAMA has the value 8, modeling the absence of assigned AMA. Contact place Air models the air, in which packets are transmitted. Transition AlloAMA sends request of AMA allocation for information exchange; the request is sent only in the presence of data packets in the output queue outbufs . Transition FreeAMA sends request of AMA release on the completion of information exchange when the output queue outbufs becomes empty. Transition send models the transmission of data packets in the presence of a valid AMA associated with slave device. Let us consider in detail the process of data packets generation represented by the left part of the model. Periodicity of the data generation is represented by the place clock ; it initializes various data packets, the number of which is given by the random function NSend () having the uniform distribution in a random interval of time given by the function Delay () . Generated packets arrive in the intermediate place data , from which they are moved in to the output queue represented by the place outbufs of the list type Transition assigns AMA to slave device on receiving receipt from master device; the check-up includes recognition of packet type (#mtype p=1), own PMA (#pma p=tpma) and the absence of available AMA (#srcama p<>8). Transition DelAMA models the release of AMA at the receipt of the corresponding packet (#mtype p=2). Note that, we have to check the slave AMA as well as write the special value 8 modeling the absence of AMA into place myAMA . . Dispatching of data into air is modeled by the transition send and is implemented only if the slave device possesses a valid AMA. Notice that, transition send listens the air and implements the transmission only in the presence of permissive label availslave . Transition models the receipt of data packets and implements their allocation to the buffer InBuf . It should be noted that, the received packet is replaced by the label availmaster providing the primary usage of the air by master device. Notice that, the model of slave device contains two contact places Air and myPMA labeled by the I/O tag. These places are further used for composition of piconets ...

Citations

... In CSPN, a random variable defined by an exponential distribution law can be used to fire the transition. Well-suited for modelling systems in which communication, resource sharing and synchronization are important, CSPN have been applied extensively in the field of communication networks[20]-[21]. ...
Research
Full-text available
This paper proposes models to compare two diagnosis architectures and justify new innovative designs in the field of railway transport. The models proposed are based on Stochastic Petri Nets and implement the concept of active diagnosis based on the notion of either High Level or Low level diagnosis. Following the theoretical stochastic validation of the models, the real case of a passenger door, which is subjected to significant stress throughout its life cycle and requires a detailed diagnosis, was studied. The first stage of the comparison concerns six validation criteria measured using Petri Net models, which represent the following indicators: reliability, maintainability, availability, error communication, time spent on analyzing alarms and complexity of the diagnosis architecture. The comparison between the two extreme diagnosis architectures, which reflect alternative distribution strategies for monitoring and fault detection and isolation functions, is discussed in relation to the aforementioned validation criteria. The main result for the practical real case is that the new architecture proposed is more reactive and more available than the current one. However, the increased reactivity of the new architecture proposed induces a decrease in reliability due to the complexity of its design.
... Simulating system CPN Tools [7] was chosen for the construction of models; it was developed in Aarhus University, Denmark and uses the language of colored Petri nets [8] for models description. In Odessa National Academy of Telecommunications named after A. S. Popov, the library of the model components was created for Ethernet, IP, MPLS, Bluetooth, E6 networks as well as the library of measuring fragments for the networks performance and QoS evaluation [6,[9][10][11][12]. ...
Article
Full-text available
Basic components of Provider Backbone Bridge (PBB) network models were constructed: PBB interior switch, PBB edge switch—with the dynamic filling up of address tables. The modeling of PBB networks was implemented. The results of simulation reveal definite imperfections of PBB technology caused by the broadcasting and sensitivity to the ageing time of the address tables' records, which complicates the guaranteeing of a given QoS. The preliminary comparison confirms definite advantages of E6 addressing before PBB.
... It is based on express-evaluations of characteristics obtained in the shortest time for new project decisions that determine the relevance of the present research. Colored Petri Nets [1] and CPN Tools [2] are successfully used for modeling Ethernet345, TCP/IP and MPLS [6], wireless Bluetooth [7] networks. Colored Petri Nets allow not only the modeling of telecommunication networks, but also the estimation of their characteristics via special measuring fragments [4,8] during the process of simulation. ...