Service interface

♦ 10/100/1000 Base-T Ethernet uplink port: 7 in GUPE7 uplink card, 3 in GUP7 uplink card, 10/100/1000M auto-select, RJ45 interface, transmit length 100m.

♦ 1000 Base-Fx Ethenrnet uplink optical port: 4 in GUP7 uplink card, 1000M speed, LC/PC connector, transmit length varies from 550m to 120km according to different optical module.

♦ 2M uplink port: 16 E1 ports.

♦ STM-1 uplink port: 2 STM-1 uplink ports in each board, SC/PC connector.

♦ EPON service port: 2 EPON office-end optical interface (PON port) in each EC2 card, SC/PC connector, max optical upstream and downstream speed reaches 1.25Gbps. Split rate of each PON port is 1:64, namely connecting 64 ONUs (remote-end unit) through splitter. Hence 128 ONUs can be connected to one EC2 card, and max 2048 ONUs can be connected to one AN5116-02 unit.

5.2.2

Management interface

♦ External management interface: 100M FE port in GUP7 or GUPE7 card, supports TCP/IP protocol, SNMPv1/v2 protocol.

♦ Internal management interface: 1000M optical/electrical port in GUP7 or GUPE7 card, supports TCP/IP protocol, SNMPv1/v2 protocol.

♦ RS232 COM port in GSWC card: for local maintenance, manage the unit via command line.

5.2.3

Other interfaces

♦ -48V DC input port.

♦ Alarm output port.

5.3

Working condition

5.3.1

Environment condition

♦ Operating temperature 0℃ ~ 50℃

♦ Storage temperature 30℃ ~ 60℃

♦ Environment humidity 10% ~ 90% (no precipitation)

♦ Atmosphere pressure 70 ~ 106 kPa

♦ No corrosion or solvent gas in atmosphere, no dust, no strong electromagnetic field interference nearby.

♦ Floor bearing >600 kg/m2

5.3.2

Power supply

♦ Voltage DC -48V (−40V ~−57V)

♦ Power consumption 650W (one full loaded subrack)

5.3.3

Unit earthing

Appendix A

Relevant Protocols and Technologies

As one of the aggregated optical broadband/narrowband network access system first developed in industry by FiberHome, AN5116-02 is designed with high port-density, high switching capability, and supports multiple routing protocols. The following will introduce main protocols and relevant technologies.

A.1

EPON

Institute of Electrical and Electronics Engineers (IEEE) started EFM (Ethernet in the First Mile) study group in the year 2000, and began to work over a brand new access technology — Ethernet PON. From then on, equipment manufacturers and telecom carriers began to learn EPON.

Simply, when combining data link layer, Ethernet and Passive Optical Network (PON) of physical layer, it comes Ethernet Passive Optical Network (EPON) communication access system. PON is an access system of one point to multipoint via optical fiber. Downstream in. PON adopts time division multiplex broadcast method, and upstream adopts time division multiple access method. Thus, it saves optical fiber and uses less optical equipments. PON is an optical network built up with passive optical devices. Lots of advantages are provided when using passive devices rather than active devices, such as of high bandwidth, highly reliable, easy maintaining, low cost, and easy for upgrading and expansion, etc.

EPON access method is of large advantages when compared with traditional access methods. If adopting traditional access method, one or more machine rooms shall be built, and there is a series of expensive cost shall be paid for machine room, such as construction cost, maintenance cost, etc. However, there is no machine room need for EPON, and the optical coupled device can replace O/E converters and switches to save multi-core fibers

than traditional access methods especially for applications of large scale.

Another benefit of EPON is its network management: the workload is very little and almost no maintenance is needed. This is also its most distinct characteristic. EPON is evolved from Ethernet technology, so it supports all maturely developed layer 2 technologies such as VLAN, etc. It also inherits intrinsic advantages of Ethernet, including its simplicity, low cost, good compatibility, flexible addressing, fairness, high speed, low latency, good stability, maintainability, etc. With SLA-based management mode, EPON can offer excellent bandwidth management. And by allocating logical link for services, it can provide users with good QoS guarantee and may well meet users’ requirements on service quality.

A.2

VLAN

VLAN can be used to divide the switch ports into different groups to establish safe and separate broadcast or multicast domain. Main purpose of creating VLAN is limiting the transmitting range of the broadcast packets and decreasing their influence. All Ethernet packets, such as unicast, broadcast, and multicast packets, as well as unknown packets are forwarded and flooded only inside the VLAN, and users don’t belong to this VLAN will not receive packets for this VLAN; that is, information for a certain VLAN will be protected from being wiretapped by users of other VLANs so as to guarantee information security, and accordingly improve network security to a certain degree.

Another advantage of VLAN is that it can change the network topology without having to physically “move” the workstations on this network into another VLAN. This means it makes the increasing, moving and relocating of network nodes flexible and convenient. This equipment provides two VLAN implementation methods: VLAN divided on the basis of the port (port-based VLAN) and VLAN divided on the basis of 802.1Q (802.1Q VLAN). 802.1Q VLAN supports the IEEE 802.1Q tag function and extends VLAN to the whole network (it requires all switches on the network support IEEE 802.1Q). And the 802.1Q VLAN untagged characteristic enables it to normally communicate with all valid switches or network cards that cannot identify a VLAN tag.

The following introduces the two VLAN implementation methods in detail.

Port-based VLAN

This VLAN implementation method establishes different broadcast domains by dividing ports into different VLANs. In a port-based VLAN, broadcast packets, multicast packets and unknown packets are all limited in the VLAN, accordingly isolating the broadcast domain.

network administrator wants to quickly and easily configure VLAN to limit the broadcast traffic on the network.

To implement the VLAN configuration more reliably, make sure that all relevant stations have been connected to the switch directly. If these stations are connected with the switch ports through a hub, switch or repeater, all unrelated stations connected to it will also be included into this VLAN.

You can create a port-based VLAN by firstly naming this VLAN and then appointing the ports in it. All the rest ports will be automatically excluded from this VLAN.

The following gives an example for the port-based VLAN.

Figure A-1 Port-based VLAN example

As shown in the figure above, the switch ports are numbered 1 to 12 from left to right. According to the port divide method, ports 1, 4, 7 and 12 form VLAN1; ports 2, 8 and 12 form VLAN2. It builds up separate broadcast domains respectively for the Sales Department and the R & D Department. And all these ports also belong to VLAN3. Port 12 is included in 3 VLANs at the same time. Ports like this are usually connected with the server; therefore, the server can receive the packets sent from two VLANs, as well as forward packets to the ports in these two VLANs. In this way, it not only divides the broadcast domains but also offers the access to the public services through the ports that are included in several VLANs simultaneously.

VLAN 2 VLAN 3 Sales Department Switch R & D Department VLAN 1

802.1Q VLAN

According to the IEEE802.1Q protocol, a switch can support up to 4094 802.1Q VLANs. 802.1Q VLAN limits data receiving and sending on the basis of IP device port. All equipments connected to a certain IP device port will become members of its VLAN, no matter it’s a single computer or all computers of a department. IEEE 802.1Q VLAN changes the previous IEEE 802.3 frame format by adding a 4-byte 802.1Q tag, i.e., VLAN Tag (see the figure below), to the end of source address (SA).

Figure A- 2 IEEE802.1Q VLAN frame structure The following gives the glossary in common use in 802.1Q VLAN.

♦ VLAN Tag: the 32-bit field in the Tagged frame header. This field comprises defined value 8100 (16 bits), User Priority (3 bits), CFI (1 bit) and VID (12 bits). The CFI stands for Canonical Format Indicator.

♦ VLAN ID (VID): the 12-bit identification in VLAN Tag to uniquely identify the VLAN.

♦ Tagged frame: the frame with a VLAN Tag.

♦ Untagged frame: the frame without a VLAN Tag, i.e. normal frames.

♦ Port VLAN ID (PVID): the identification used to associate a VLAN with a port. For example, the port with PVID 1 will forward all its input frames to the VLAN with VID 1.

♦ Untagged port: the ports that join in a VLAN with Untagged mode. These ports only send untagged frames; that is, frames sent from them are all untagged.

♦ Tagged port: the ports that join in a VLAN with tagged mode. These ports only send tagged frames; that is, frames sent from these ports are all tagged.

8 6 6 2 2 2 Variable 4

Preamble

SFD DA SA 8100

Type

Length Data FCS

User Priority (3 bits) CFI (1 bit) VID (12 bits) VLAN Tag

A.3

STP

In data transmission, redundant links are required as backups in the case of break of primary link to avoid network paralysis. However, redundant links on network cause the potential exists for data forwarding circulation and accordingly cause endless loops. The switch will automatically take the optimal path and disable the other redundant paths to avoid the creation of loops. On the other hand, it will establish redundant paths in the event of break of primary link to avoid paralysis of the whole network.

The protocol for exchanging information among bridges is referred to as STP. With its algorithm, the bridges can dynamically create a loop-free subset of the topology, or a tree, and at the same time possess enough connectivity so that if physically possible, only one path exists between each two LANs. STP reconfigures the network and reroutes data paths by activating the appropriate standby path.

The basic concept of STP algorithm is that the bridges create the spanning tree by exchanging special messages among them. In IEEE802.1D, this special message is called BPDU.

There are two Spanning Tree Algorithm (STA) Operation Levels: bridge level and port level. At the bridge level, STA counts Bridge Identifier for each switch and specifies the Root Bridge and Designated Bridges. As for the port level, STA specifies the Root Port and Designated Ports.

A.4

Port trunking

Port trunking is a method of binding multiple ports of lower bandwidth as a single link with greater bandwidth to balance the link traffic load via several ports and thus avoid link congestion. It is like the supermarket setting multiple checkout counters to obviate longtime queue up of consumers due to too few checkout counters. AN5116-02 equipment supports port trunking function. This means that it can connect multiple physical ports as a single logical port to achieve a greater bandwidth. In addition, it enhances the reliability of the connection between equipments. If one of the ports in the trunk group fails, the traffic on that port is automatically forwarded via the other ports in the trunk group, effectively assuring continuity of the connection.

All the ports included in a trunk are treated as a single port and one of them is appointed as the master port. All the ports in a trunk operate in just the same mode, and therefore all configuration for the master port will be applied to all the other ports in this trunk, so all you have to do is just configuring the master port. In addition, all the ports in a trunk are regarded as a single port for such functions as VLAN, STP, etc.; that is, all operations are only required on the master port.

A.5

Multicast

Multimedia services over Internet, such as streaming media, videoconference, video on demand, etc., have become an important part of information transmission. The point-to- point unicast method is not suitable for delivery of this kind of services, i.e., one- transmitter/many-receiver. In unicast mode, the server must provide each receiver with an IP message copy of the same contents, and messages of the same contents are transferred over the network repeatedly, occupying a lot of resources. Although IP broadcast permits one host to send one IP message to all hosts on the same network, not all of them actually need these messages, so this mode also wastes network resources. Multicast is introduced as required to solve these problems. It offers the host a method to deliver messages to a specified group of recipients, as shown in the figure below. In 1989, IETF passed the RFC1112 standard, which defines the multicast method over the Internet.

Figure A- 3 Multicast

Non-multicast

Host Host Client Server

Non-multicasttransmission

Host Host Client Server

IGMP Snooping

IGMP snooping allows a switch to "listen in" on the IGMP conversation between hosts and multicast servers. When a Switch hears an IGMP report from a host for a given multicast group, the switch adds the host's port number to the IGMP list for that group. And, when the switch hears an IGMP leaves, it removes the host's port from the IGMP list for that group.

IGMP snooping function manages layer 2 multicast traffic on a switch. This function provides the switch with the ability to control the multicast traffic so that it travels only to those destinations that require it and thus reduces the amount of broadcast traffic and saves the network bandwidth. When the switch starts IGMP snooping function, it creates a multicast forward table for each VLAN. And when the switch receives a report of join report from a host, it will automatically add the corresponding port number into the relevant multicast forward table. This function is very useful for video multicast applications: instead of delivering an individual copy to every interested recipient, the server replicates the video stream layer upon layer by using IP Multicast and accordingly lightens the network burden.

Multicast is a network technology that allows one or more senders (multicast source) to deliver a single stream of information to more than one recipient (at one time, simultaneously). The multicast source sends packets to a given multicast group, and only the destinations that belong to this group can receive these packets. IP Multicast can greatly save network bandwidth, for only a single stream of information is delivered over any link on the whole network regardless of the number of the recipient. It improves the data forwarding efficiency and reduces the possibility of backbone congestion. The multicast group does not have any physical or geographical boundaries; that is, the hosts can be located anywhere on the Internet with support from the multicast router.

Precondition of IP Multicast implementation

To implement IP multicast, the multicast source and recipients and the underlying network between them all must support IP Multicast, including the following aspects:

♦ The network interface of the host allows for Multicast;

♦ A set of group management protocol used for join, leave and query, i.e. IGMP (v1, v2);

♦ A set of IP address allocation policy, and can map layer 3 IP multicast addresses into layer 2 MAC addresses;

♦ IP Multicast application software;

♦ All routers, hubs, switches, TCP/IP stacks and firewalls between the multicast source and recipients support IP Multicast,

Definition of IP Multicast Address

Multicast communication needs two types of addresses: IP multicast address and Ethernet multicast address. IP multicast address is used to identify a multicast group. Ethernet multicast address is required because all IP packets are encapsulated into Ethernet frames. The host must receive both unicast traffic and multicast traffic to make IP multicast operate normally, and this means it needs multiple multicast IP addresses and Ethernet addresses. In IPv4, the Class D address space, from 224.0.0.0 to 239.255.255.255, has been assigned for IP multicast. Moreover, the Class D address is divided into Reserved Link Local Addresses, Globally Scoped Address and Limited Scope Addresses, whose meanings are listed as follows:

♦ Reserved Link Local Addresses: 224.0.0.0 ～ 224.0.0.255, used on a particular LAN segment. A router should never forward packets with these addresses;

♦ Globally Scoped Address: 224.0.1.0～238.255.255.255, can be used to multicast data between organizations and across the Internet;

♦ Limited Scope Addresses: 239.0.0.0～239.255.255.255, constrained to a local group or organization, used to define multicast boundaries;

The last 28 bits of the Class D address have no changes in the structure, i.e. without differentiation between network ID and host ID. An arbitrary collection of hosts that respond to a certain IP multicast address form a multicast group. A multicast group can span several networks. The members of a multicast group are dynamic; a host can join or leave a certain multicast group using IGMP. Because the upper 5 bits of the IP multicast address are dropped in this mapping, the resulting address is not unique. In fact, 32 different multicast group IDs all map to the same Ethernet address.

IP Multicast Protocols

IP Multicast Protocols mainly include IGMP and IP Routing protocol.

IGMP

A host uses IGMP to inform the subnet multicast router and apply for joining an IP multicast group. A router applies IGMP to discover whether any host on the local subnet belongs to a certain IP multicast group.

Joining a Multicast Group

When a host wants to join an IP multicast group, it sends a Host Membership Report message to the IGMP router of the IP subnet where it is located, and at the same time prepares its IP module to receive the packets for this IP multicast group. If this host is the first one on its IP subnet that joined this IP multicast group, the IGMP router will be added into the multicast distribution tree through routing message exchange.

Leaving a Multicast Group

For IGMP version 1, if a host wants to leave a certain IGMP group, it just silently quits the group. The IGMP routers periodically (per 120 seconds, for instance) send Host Membership Query messages to inquire the group address (224.0.0.1) of all hosts on this IP subnet. If no member is in a certain IGMP group on a certain IP subnet, the IGMP router will not forward packets for this multicast group to this IP subnet. Simultaneously, the related IGMP router will be removed from the relevant multicast distribution tree through routing message exchange. This leaving silently without informing anybody causes latency in the IGMP router’s awareness of the event that there is no member on the IP subnet.

While in IGMP version 2, if a host wants to leave a certain IGMP group, it notifies the IGMP router of the IP subnet, which immediately inquires all the IGMP groups on this IP

A.6

V5 protocol

V5 interface

The subscriber line has a length of up to almost 5 km. When the distance between the Private Branch exchange (PBX) and the subscriber exceeds this limit, extra PBXs must be adopted by the telecom office to set up a new office, usually referred to as C5 office. The telecom office also has to bear all sorts of administrative costs, including the extra equipments, manpower, materials, etc. To reduce these costs, many telecom equipment suppliers have launched the remote module. Its function is gathering up the PBX side subscriber line signaling via the 2M trunk lines and transmitting it to the remote module, voice channel cards of which will transmit it along. For the remote modules from different suppliers comply with different protocols, the interworking of this mode is weak. For this reason, the V5 protocol is brought up. It specifies the communication protocol between the PBX and the remote module, calls the reference point of the interface between the PBX and the remote module V5, and names the protocol in that course V5 protocol. V5 protocol includes V5.1 protocol without line concentration function and V5.2 protocol with line