Seamless switching in mobile multimedia networks

Absrtact: The third generation mobile communication faces an important challenge is to seamlessly integrate the fixed and mobile network multimedia services. For mobile users, the network supports the business of image, multimedia, data and different service levels of voice services. In order to meet the above business requirements, 3G system must have a rich performance. While homogeneous networks using ATM technology today support many users, such a network structure cannot be the ultimate solution (at least from the perspective of academia, network equipment manufacturers). The rapid development of high-speed devices in the Ethernet family has partially replaced ATM. At the same time, based on the explosive growth of Internet services, it has been ensured that IP will remain the next generation System Network layer protocol. This paper discusses the problem of IP network in support of mobile service and the analysis of switching technology.

1. Introduction

In the future network structure, the mobile network is composed of a series of layers with their own characteristics and overlapping. The satellite, the macro community, the Microcell and the Pico community cover different geographical regions. and support the mobile terminals at different speeds. Some layers belong to private operations, and some layers belong to public operations. The rate of the Pico plot exceeds 25Mbps. The coverage area is the building area, the Microcell can reach the maximum 2Mbps, the densely populated town area, the macro plot rate is hundreds of kbps, the coverage area is wide, the satellite provides the mobile terminal speed to reach 144Kbps above, covers the entire continent.

Traditional communication devices can only be connected to a single type of network. The global interconnection network, which is based on terrestrial communications (a vast and complex communication network and the Internet that is currently made of cables, optical fibres, copper cables), provides a communication link for fixed systems. and the gateway between fixed network and mobile network, such as GSM, extends this global interconnection to mobile terminals. such as: terminal identification, registration authorization, terminal mobility management, terminal roaming business re-election route management. The re-election routing mechanism is also called switching, which is the key part of the mobile network structure.

The idea of using a multi-format mobile phone has been accepted by the Global Mobile Communication System (UMTS). Currently provided by GSM to provide a wide range of coverage, using WCDMA macro cell and Microcell to provide a relatively high broadband area as a supplement. GSM and WCDMA layer switching will be supported in the network recommended in UMTS. (Computer science)

The fast and data-free handoff among different network types is one of the key problems in creating seamless mobile multimedia network.

2. Basic Concepts

The network is made up of some interconnected members. The internet is composed of a series of interconnected networks and subnets. Each subnet has a network address as an identifier to differentiate it from other subnets.

A series of interconnected host nodes form subnets. Each node is identified by an address (node number) that is unique on the subnet. The network number and the node number are the unique identities of the nodes in the Internet. Since the node is static, this unique identification is used for the permanent address of the node.

Subnets are connected to each other through routers. In order for the router to work correctly, a unified authority manages and assigns subnet addresses. An important management principle of this unified authority is that duplicate addresses are not allowed to occur. A packet on the internet that has a source and destination address attached to the message header. The router detects the destination subnet address in the packet at the input end, and then determines that one output is the next route to the destination of the packet. Typically, there are multiple routes between the given source and destination nodes. The routing table is designed to allow the router to select the next path where the data reaches the end point. Therefore, the node's network number is used by routers to determine the subnet address in the Internet, and the node number is used to identify the nodes in the subnet.

When a node moves through the Internet, the permanent address of the node cannot identify the path to the node. Using the concept of mobile networks, mobile hosts (MH) connect to the Internet via a network connection point (NPOA). The permanent address of the mobile node remains unchanged, and NPOA changes as the node moves from one subnet to another, which requires the routing information of the mobile host to maintain the relationship between the permanent address and the NPOA.

The above concepts can also be used in other networks. For example, circuit switching networks (such as telephones and ATM networks), switching replaces the role of the router, and the circuit identification replaces the source and destination addresses.

3. Basic switching type

When the IP network supports the mobile service, it is confronted with a series of problems, which are discussed from two aspects of horizontal and vertical switching.

In a circuit switching network, a switching process is described as moving an end-to-end communication device from one base station to another base station. The base station provides a network access point (NPOA) for the mobile terminal. In the network, the route is updated to recover from a network that has failed, congested, or modified the topology. By using NPOA, one function of switching is to track mobile terminals, while subsequent routing updates redirect the business to another more appropriate route.

Because the network has different hierarchies, we refer to the update in the same network as the "horizontal switch", and the updates to the middle of the network are called "vertical transitions". This enables the concepts of switching and routing updates to be consistent.

Latency and data loss are key factors in seamlessly switching between two. Horizontal switching typically occurs when a mobile device roams between two or more cells. Low latency can be guaranteed when switching between adjacent small intervals in the same network. In some special environments, such as the high transmission rate between tiny cells or the high density of mobile devices, horizontal switching can meet the requirement of no data packet loss.

Generally speaking, the network coverage area with high transmission rate is smaller than the network coverage area with low transmission rate. Therefore, the vertical switch to a high transmission rate network, you can ensure that the packet is not lost, such as: from the micro-area switch to the small cell. The premise is that for mobile devices or use these two networks, if the transmission rate of high network layer congestion, switching will not occur, the packet will not be lost.

If the current network becomes unavailable, such as the user roaming the scope of the macro community coverage, the mobile terminal regardless of whether the simultaneous access to both networks will occur vertical switching. When the network connection is temporarily interrupted, or the new network layer transmits at a lower rate than the previous network layer. The packet is lost at this time.

Switching between different network layers also results in a change in the quality of service provided by the network to the mobile terminal. For example, switch from Wireless LAN (2Mbps) to GSM (9.6kbps). There are two types of scenarios that support bandwidth changes. First, the business based on mobile terminals supports different service quality. The disadvantage of this approach is that mobile terminals cannot support "standard" business. The second approach uses a network proxy or packet filter to support the "standard" service used by the same network's terminals.

Application of 3.1 Vertical switching technology in IP network

In a data network, the general way to support mobility is to use a switch or router to maintain the relationship between the permanent address of the mobile device and the NPOA. This approach is used in IP networks.

In a non mobile IP network, the nodes in the network are fixed. The IP address uniquely identifies the node and the subnet to which it is connected. Packets are sent to the corresponding IP subnet based on the destination address packet. If a node is a removable host, when it roams to a new subnet, the IP address and the old subnet associated with the node and the existing routing relationships are broken, so that the packet cannot be sent to the new NPOA as the previous routing connection.

The basic scheme used in mobile IP networks is to extend the IP protocol, allowing transparent IP packets to be routed to the mobile host (MH). When MH is connected to the home network (HA), when MH roams to another network, it needs a forwarding address (COA) to identify the address that MH is currently connected to the Internet. The COA can be the address of a foreign agent in a local village, or the address of the current connection obtained by MH (e.g., using DHCP). The COA is registered to the home agent (HA) of MH's home network.

The asymmetric routing mechanism used in mobile IP (also known as "triangular path"). The route from MH to the Communication Host (CH) is direct, while the packets from CH to MH are intercepted by Ha in the MH home network (the subnet identified by the MH IP address). HA uses COA to transmit packets forward to MH. Therefore, the path of packet delivery is significantly longer than the optimal path.

The protocol of the current IP network is restricted by the address range for Ipv4,ipv4. The IPV6 are being developed to overcome these limitations. All two versions support the mobility of IP networks, and IPV6 supports routing optimization.

The IPV6-supported routing optimization surface saves connection tables for all nodes (fixed or moving) to cache the connection relationships between mobile nodes. This connection actually represents the connection between the home address and the COA. When the packet is sent to the IPV6 destination, the node checks the cached connection table to determine the destination address of the packet. If found in the relational table, the node sends the packet directly to the COA address in the relational table. When MH roams to another network, MH must send the new COA to the associated communications host. With a cached connection table, packets can be delivered directly without passing through HA, which makes routing updates more efficient.

Application of 3.2 Horizontal switching technology in IP network

There are two main technologies that support horizontal switching in IP networks: IP multicast and ARP switching.

IP multicast: IPV4 has three types of addresses: Single transmission, broadcast and multicast. A single routing address is used to transfer packets to a single destination. Broadcast addresses are used to transfer packets to the entire network. Multicast addresses are used to transfer packets to a set of hosts that are configured as members of a multicast group and can span different networks.

Multicast is mainly applied to the Internet, such as: Video conferencing. The same information is sent to different hosts. With a multicast address, the same information can be replicated by the server and sent to each host. With a multicast address, the server only has to send out a packet, and the multicast router consists of a multicast group allocation tree, which is routed forward through the allocation tree until the next multicast router; When the allocation tree splits, the router automatically replicates the packet.

Multicast is also used in Daedalus engineering to support switching between wireless LAN communities. In this project, mobile IP provides a single connection point for mobility, and the COA is a multicast address that does not change.

The current mobile base station and several surrounding base stations represent the multicast groups in the move. They all receive packets from the mobile terminal and cache the packets, but only one of the current base stations transmits the packets to the mobile terminal.

The mobile terminal predicts the occurrence of the switch by measuring the signal strength of the received nearby base station. When a mobile terminal initiates a switch, it sends the message to the original and new base station, and the latter begins sending the packet to the data terminal.

Multicast, because it provides a mechanism for fast routing updates, has the advantage that routing update messages are not sent to the home agent (HA). One of its drawbacks is that the purpose of a multicast address is to transfer packets to a mobile terminal on a different network. In this case, a mobile terminal has only one multicast address.

ARP Toggle: ARP is an underground resolution protocol applied to the same subnet. Map the IP address of a node to the corresponding hardware address (MAC address). The hardware address is used to transfer packets.

In an IP network, when a packet arrives at the network layer, the router sends an ARP request to find the physical address of the device in the packet. ARP answers are not necessarily from the destination itself, but from the network agent who knows how to send the business to the destination.

In mobile communication, the base station acts as the ARP proxy for MH in the cell, and it sends the data packages to MH. Through the wireless link, the base station sends a signal containing its own address and caches the latest packets to MH.

When MH receives more signals from a base station than the current base station sends, MH initiates a switch, and the switch process is as follows:

1 MH sends a greeting message to the new base station, and the message contains the MH address and the original base station address.

2 The new base station creates a routing portal for MH and responds with a greeting confirming message.

3 The new base station sends a notification message with the address of the new base station through the wired link to the original base station.

4 The original base station deletes the MH inlet and sends the cached data packet to the new base station by confirming the message to the original base station.

5 The new base station broadcasts a redirected message to the associated routers and hosts, updating their ARP cache information.

ARP switching requires the completion of 10ms in the case of a limited packet loss. This requirement is related to the size of the buffer. Because the ARP protocol can only be restricted to one subnet, this switching technique can only be used for switching between base stations within the same subnet.

4. Concluding remarks

In this paper, the seamless handoff technology in Mobile multimedia network is analyzed. The switching speed of the mobile terminal is faster and the data packet is not lost and the requirement of seamless switching is achieved. In the vertical switching between different networks, when the switching occurs in the network with low transmission rate, the switching speed is faster and the packet is guaranteed to be free of loss. Switching in the opposite direction can lead to loss of packets. This requires that multimedia services not only know the current level of data service available, but also require the ability to perceive service level changes and adjust QoS accordingly to ensure that packets are not lost when a switch occurs.