**Ring and node architecture**

01S can be mesh network or ring network. The 01S ring network is shown in figure below.

OISN: Optical integrative switching ring node OISR: Optical integrative switching ring network

BS: WiMAX Base Station

The system has n OIS nodes in an unidirectional OIS ring. The OIS ring can be a backbone of wireless metropolitan area network (MAN) that interconnects a number of WiMAX base stations.

Fig. 4. Architecture of OIS ring network

The optical integrative switching ring network has nxm+l wavelengths. The total data transmitters number in the 01S ring network is nxm corresponding nxm WDM wavelengths. Each node has m wavelengths which are referred as the home wavelengths of the node. The (nxm+l)th wavelength is the control channel for the wavelengths in the set and it carries control frames. The control frames implement the signaling necessary for 01S. Each 01S node in the ring has m data transmitters each fixed-tuned to one of them home wavelengths, and m tunable receivers. These m pairs of transceivers are used for transmitting and receiving bursts . A node can only transmit bursts on its home wavelengths . A free receiver can tune to receive a burst arriving on any wavelength in the ring. A WTFM (wavelength tunable filter module) was installed at the destination node to drop any m data wavelengths and the control wavelength. The network scheme has the

feature of transmitting wavelengths collision free, traffic collision free when switching at the midst nodes, traffic collision free from the source node. The core WDM network utilizes the W-token (token supporting WiMAX) signaling scheme to resolve receiver collisions at the receivers. In the W-token access protocol, the token can be captured according to the service class. The 01S node is equipped with a control module which performs its functions based on the information each control frame carries around the ring. Each node has its own slot into which it can write information during transmission. The control frames on the m control wavelengths travel around the ring in a synchronous manner.

CPL: coupler WTFM: wavelength tunable filter module RxC: the receiver of control channel TxC: the transmitter of control channel Rx: the receiver of traffic channel Tx: the transmitter of traffic channel

Fig. 5. The Node Architecture of OIS ring network when m=2

Each 01S node serves a number of WiMAX base stations. In the direction from the WiMAX base stations to the ring, the 01S node collects and buffers electronically data packets, transmitted by users over the WiMAX base stations. Buffered packets are subsequently grouped in different burst according to their classes and destination. A burst can be of any size between a minimum and maximum value. Bursts travel as optical signals along the ring, without undergoing any electro-optic conversion at intermediate nodes. In the other direction from the ring to the WiMAX base stations, an 01S node terminates optical bursts destined to itself, electronically processes the data packets contained therein, and delivers them to users in its attached WiMAX base stations.

A WTFM was installed at each destination node before m data receivers and the control channel receiver to drop any m data wavelengths and the control channel wavelength. The architecture of the node when m=2 is shown as Fig.2. When the node receives the wavelengths from the ring, the two data wavelengths and the control channel wavelength was dropped by the WTFM. The other wavelengths were coupled with the two home data

wavelengths (Al and A2) which transmit the burst of the node as well as the control channel wavelength and transmit to the OIS ring network.

In OIS ring networks technology supporting WiMAX, there are four different classes of service traffic. The 0 class of traffic (Oass 0) can support Unsolicited Grant Service (type 0) in WiMAX; The class 1 traffic can support Real-Time Variable Rate Service (type 1) and Extended Real-Time Variable Rate Service (type 4) in WiMAX; The class 2 traffic can support Non-Real-Time Variable Rate service (type 2) in WiMAX; The class 3 traffic can support Best Efforts Service (type 3) in WiMAX. For service flows, the class parameter should be used when determining precedence in request service and the lower numbers indicate preemptive priority. Bursts with different classes are served by using W-token access protocol. In the W-token access protocol, token can be released and captured according to the service class. In the OIS network, token0 and tokenl are distributed for every receiving port to indicate the receiving port is unoccupied. Token0 can support class0 service and tokenl can support other service in OIS network.

## **Control wavelength operation**

The control channel wavelength is used for the transmission of control signal. In a ring with n nodes, nx4xm control slots, 4xm slots corresponding to m transmitters of each node, are grouped together in a control frame which continuously circulates around the ring. Four control slots includes token0, tokenl, answer frame (ACK) and a control frame (CF) corresponding to each transmitter of each node.

Each node is the owner of 4xm control slots in each control frame. Each control slot contains several fields. Each token0 control slot includes fields for the destination address, the token serial number, class of traffic is 0, the offset, the transmitter wavelength of exist class0 service, the time slot of exist class0 service, the time slot of reserving class0 service and so on. Each tokenl control slot includes fields for the destination address, the token serial number, class of traffic, the offset, the transmitter wavelength, the time slot of exist class0 service, the time slot of reserving class0 service and so on. Each answer frame (ACK) includes fields for the destination address and the source address, the token0 serial number, class of traffic, the offset, flag, the transmitter wavelength and the size of data, the time slot of exist class0 service, the time slot of reserving class0 service and so on. Each CF control slot includes fields for the destination address, the tokenl serial number, class of traffic, the offset, flags, the transmitter wavelength, the burst size the time slot of exist class0 service, the time slot of reserving class0 service and so on.

The value of traffic class is as follows: 0 is corresponding to class 0 traffic, 1 is corresponding to class 1 traffic 2 is corresponding to class 2 traffic, 3 is corresponding to class 3 traffic, and 9 is corresponding to no traffic. The offset 1 value is the processing times at intermediate nodes and the offset 2 value is the time of WTFM to switch to receive different wavelength. The offset value is the sum of the offset 1 and the offset 2. The offset time doesn't include the transmitting time from source to destination.

#### **1.2.2.2 The W-token access protocol in 01S network**

The protocol uses tokens to resolve receiver collisions at the receivers. Tokens are used for all classes. Every node has m token0 and m tokenl corresponding to m receivers of this node circulating around the ring. If a source wants to transmit a burst to a particular

destination, it has to catch the token for that destination according to classes of service. To accommodate the extant PCM network (including SDH), 125µs is adopted as data period for receiving port. A token may be either available or in use. When the sending node catches token0, the time slot of receiving port will be reserved. If the time slot of receiving port reserved successful, the connection of class0 will be established. Then the class0 occupy the same time slot in every period of 125µs. This message of occupying time slot will be write in token. The else service will keep away from this time slot when they catch token.

Since only the node that has possession of the token can transmit a burst to the appropriate destination in its time slot, the receiver of the destination can only receive a single burst at a time, and therefore the W-token protocol is a receiver collision-free protocol.

The maximum bandwidth of each class can be defined by the network. The service bandwidth which has a lower priority can be taken up by the higher priority; however, the defined maximum bandwidth can not be exceeded. Processes of each class service sending as below:

## **The processes of sending class O service**

At the sending node, the node checks each received control frame. If the node detects an available token0, it deletes token0 from control frame, and puts token0 into its own FIFO token queue. If there is no class 0 service, the node will release token0 to the next node, otherwise, it will examine the receiving port time slot pool, and detect whether there are available time slots of the receiving port which is limited by the network (The time slot which is not occupied by class 0 can be regard as an available time slot. When the time slot is occupied by lower priority service, it can be grabbed by higher ones. However, the class 0 bandwidth can not be wider than the maximum class 0 bandwidth which is defined by network. According to the network resources, network can define the maximum class 0 service bandwidth of the whole network. The maximum connections bandwidth between each pair of nodes can be defined. Idle available time slot means it is longer than data required). If there is none, the node will release token0 to the next node, else, it will examine local sending time slots, and see whether there are idle available sending time slots (If the total bandwidth including the bandwidth which is required by the sending data is not larger than the class 0 service maximum bandwidth, then it has idle sending time slots, or else it has no idle sending time slots. Idle available sending time slot means the idle time slot is larger than required time slot of data sending),if there is no idle available time slots, the node will release token0 to the next node, otherwise it will reserve sending time slots, and write reserving message into token0, then release token0 to the next node.

When the receiving node reveives token0, it will send the controlling order to WTFM to receive the wavelength which carrying data during the time slot and to receive class 0 service data according to the reserved time slots. Class 0 service coming from different source nodes is received according to the time slots in the frame with the period of 125µs. The frame period 125µs is defined for receiving ports, and different sending nodes occupy different time slots of receiving port. Because the wavelengths of different nodes are different, the WTFM may be used at the receiving ports. Only one OIS class 0 service data packet from one sending node is sent in each frame with the period of 125µs (Multiple El frames may be included in the period of 125µs.). Class 0 service supports WiMAX type0

service. When the receiving port was reserved to receive class0 service, the receiving node will send an ACK message to the source node and the receiving node writes arrangement message to token0 and releases it to the next node. When receiving tokenl and CF, it writes into the time slots information, and releases it to the next node. After the source node receives ACK which indicates that reserving time slots is successful, it starts to send class 0 service. When finished sending data and receiving token0, the source node write the releasing message into token0 and releases it to the next node. When received the token0 including the releasing message, the receiving node sends controlling order to the WTFM and releases the connection. It writes this time slot releasing message to token0 and ACK and release to the next node. When receiving tokenl and CF, the time slot releasing information is updated and tokenl and CF released to the next node.

#### **The processes of sending class 1 service**

At the sending node, it checks each received controlling frame. When it detects an available tokenl, then it deletes tokenl from the controlling frame, and puts tokenl into its own FIFO token queue. If there is no classl service, the node will check whether there is class2 or class3 service. If there is classl service, it will check whether there are idle available time slots at the receiving port (Idle time slots of the receiving port here is the time slots not occupied by class 0 or class 1. When the time slot is occupied by lower priority service, it can be grabbed by higher ones. However, the maximum class 1 bandwidth defined by network should not be exceeded. According to the network resources, the network could define the maximum class 1 service bandwidth of the whole network, and also it could define the maximum connection bandwidth between each pair of nodes. Idle available time slot means it is longer than burst packets required), if there is none, the node will check whether there is class2 or class3 service. If there are idle available time slots, then it checks whether there are local idle available sending time slots limited by the network (If the sum of the bandwidth of the existed class 1 service and the required bandwidth of sending time slots is not larger than the maximum bandwidth of class 1 service defined by the network, then the node will consider that there are idle sending time slots, otherwise, there are none. Idle available sending time slot means that it is larger than the burst packet sending time slot, and the time slot doesn't conflict with the idle time slot of the receiving port),if there are none, the node will go on to check whether there is class2 or class3 service. If there are idle available sending time slots, the message of occupying time slots will be written into tokenl, and then to check whether there is class 3 service. At the same time, the sending node writes the message about the occupied time slots into the control message CF and sends it to control frame. And then sends classl data after the offset time. Every time the node captures tokenl, it could send multiple OIS classl service burst packets in different frames whose period is 125µs (each class 1 service burst packet can contain multiple MPEG frames etc). Only one OIS classl service burst packet is sent in each frame with the period of 125µs. The total bandwidth should not larger than the maximum bandwidth defined by the network. Receiving CF, the receiving node will send control order to the WTFM and receive classl service data according to the reserved timeslot.

#### **The processes of sending class2 service**

Check whether there is class2 service, and if there is none, it will check whether there is class3 service. If there is class2 service, it will check whether there are idle available time

slots in the receiving port (The idle available time slot here means the time slots which are not occupied by class0, classl and class2. The service can grab the time slots of service with lower priority. The total bandwidth of class2 can not be larger than the maximum class2 service bandwidth defined by the network. According to the network resources, the network can define the maximum class2 service bandwidth of the whole network. It can also define the maximum connection bandwidth between each pair of nodes. Idle available time slot means it is longer than the burst packets required), if there is none, the node will go on to check whether there is class3 service. If there are idle available time slots, it will check whether there are local idle available sending time slots limited by the network (If the sum of the bandwidth of existing class2 service and the bandwidth of sending time slots required is not larger than the class2 service maximum bandwidth defined by the network, then the node will consider that there are idle sending time slots, otherwise it considers there are none. Idle available sending time slot means it is longer than that required by the burst packet, and the time slot does not conflict with the idle time slot in the receiving port), if there is none, it will go to check whether there is class3 service. If there are idle available sending time slots, it will write the message about time slot occupying into tokenl, then go to check if there is classs3 service. At the same time, the occupied time slot message is written into the control message CF and then CF was sent. The class2 data was sent after the offset time. When the node captures tokenl, it can send multiple OIS class2 service burst packets in different frames with the period of 125µs. When receiving CF, the receiving node sends control order to WTFM and then receives class2 service data according to the reserved time slots.

#### **The processes of sending class3 service**

The node checks whether there are idle available time slots at the receiving port (Idle available time slot at the receiving port is the time slot which is not occupied by class0, 1, 2 and 3. However, the class 3 bandwidth can not be wider than the maximum class 3 bandwidth which is defined by the network. According to the resources, the network can define the maximum class3 bandwidth of the whole network. The maximum connections bandwidth between each pair of nodes can be defined. Idle available time slot means it is longer than burst packets required), if there is none, tokenl will be released to the next node. Otherwise, it will check whether there are local idle available time slots limited by the network. (If the sum of the bandwidth of existing class3 service and the required bandwidth of sending time slots is not larger than the maximum bandwidth of class 3 service defined by the network, then the node will consider that there are idle sending time slots, otherwise it considers there are none. Idle available sending time slot means it is longer than that required by the burst packet, and the time slot does not conflict with the idle time slot in the receiving port), if there is none, the node will release tokenl to the next node. If there are idle available sending time slots, it will write the message about time slot occupied into tokenl, then release tokenl to the next node. At the same time, the occupied time slot message is written into the control message CF and then CF was sent. The class3 data was sent after the offset time. When the node captures tokenl, it can send multiple OIS class3 service burst packets in different frames with the period of 125µs. When receiving CF, the receiving node sends control order to WTFM and then receives class3 service data according to the reserved time slots. After the receiving node receives tokenl, it will write the message about the time slots occupied by class 0 service into tokenl.
