wireless and mobile data

14 - Wireless and Mobile Data CharacteristicsAlthough all wireless networks are different they do share some common characteristics.. Finally, tying the pieces together we have the backb

1 Wireless vs Mobile

2 Wireless Characteristics

3 Generic Wireless Network

4 Micro- and Macro Mobility

5 Mobile IP

6 Connectivity

7 Packet Switched vs Circuit Switched

8 Wireless Data Communications Systems

9 Wireless LAN and MAN

10 Mobility and Bitrate map

Ch 14 - Wireless and Mobile Data Characteristics

Many people use the terms wireless and mobile as synonyms This is not strictly true In this table we try to illustrate the differences between the terms wireless and mobile

Those users that are not wireless and not mobile are easy to come up with This is what you can call "traditional data communications" and consists of conventional LAN, MAN and WAN technology, such as ethernet, FDDI or X.25

At the other extreme, there are users that are both wireless and mobile So here we can have taxis using radio modems, field service engineers (recall that Mobitex's first application was for Telia's field service engineers), transport and public safety users

What about the users that are using wireless technology but are not mobile? These are the traditional "telemetry" applications that were mentioned in chapter 13 These include

parking or gas meters and coke machines

The last set of users, those that are mobile but not wireless are a little harder to define Imagine the following scenario, I sit in my office in Gothenburg and work on my laptop PC connected to my local network Then I pick up my laptop, jump on a plane and fly to Stockholm There I plug my PC into the LAN in Stockholm and carry on working as if I was

on my home network In this case I'm using wired technology - the two Local Area

Networks, but in some sense I'm mobile because I move around the country This type

of access is called Nomadic Computing An example of technology that allows this kind of movement is Mobile IP, which will be explained in more detail later

Because wireless data uses an unguided medium it suffers from certain problems that do not affect its wired counterpart One must remember that a wireless user is constrained by the network coverage If the user moves to an area without coverage, then the user's

terminal becomes useless Although network operators try to minimize these "blackspots" they are inevitable, especially for new networks that are in the process of being deployed

Multi-path fading, log-normal fading, inter-symbol interference and attenuation are all

characteristics of the radio medium Coupled with the limited amount of frequency spectrum available to wireless applications these attributes mean that the raw bit error rate is greater than that found in guided mediums This means that greater overheads are required to

maintain acceptable error rates, thus reducing the amount of user data that can be sent per second

So, wireless systems almost always offer lower bit rates than their wired alternatives

Wireless systems also tend to have a higher latency than wired systems For most protocols this is not a big issue, but is something that should be addressed when setting protocol parameters (e.g., timeouts should generally be longer due to the slower response times)

Ch 14 - Wireless and Mobile Data Characteristics

Although all wireless networks are different they do share some common characteristics Here we try and illustrate those parts, which most systems share

Most wireless applications consist of a user wishing to communicate with a server at some remote location The user's terminal is connected to the wireless network via a radio link to a base station The server is connected to a network, be it the Internet or some corporate Intranet Now, each base station is connected to something called a serving node The

terminology differs depending on which mobile network you look at, but the idea is almost always the same Note that one serving node can be attached to several base stations On the other side of the diagram we have the network attached to some gateway Finally, tying the pieces together we have the backbone network to which both the gateway and the serving node are connected Note that the backbone can be connected to several serving nodes and several gateways Using this network a wireless user can communicate with a remote server

Other things are also attached to the network backbone For example, there is some kind of subscription register to store information on the mobile users There is also some operations and maintenance equipment to allow network monitoring, charging etc

Micro and macro mobility are terms used to describe different methods of dealing with mobility management Micro mobility techniques are used when mobiles move between individual base stations - micro means small and so micro mobility techniques are used for small moves

Macro means big, so macro mobility techniques are used for large moves between different networks So, for example, when you take your GSM telephone from your home network in Sweden, and move to a new network in a different country, for example in England, you are moving to a new network and therefore the mobility management is solved using macro mobility techniques

Now, you may have spotted there is a grey area in this definition If micro mobility is

moving between base stations and macro mobility is moving between networks, what term

do we use for moving between different serving nodes?

There is no right and wrong answer here; it depends on the techniques implemented in the network For example, in a GSM network the same techniques are used when mobiles move between base stations and between serving nodes So, GSM solves the problem of moving between serving nodes with micro mobility techniques Conversely, PDC, the Japanese mobile telephone standard, uses macro mobility techniques for mobiles moving between serving nodes, so moving from one serving node to another and moving from one network to another both use the same technique It's simply a question of how the network is

standardized

Ch 14 - Wireless and Mobile Data Characteristics

What is Mobile IP? Well, it's a macro mobility technique that is it can be used to allow movement between two interconnected networks In this example we are using Mobile IP to allow a portable user to move between two different LANs interconnected by the Internet

Imagine this scenario You normally work in you lovely air-conditioned office in Dallas One day your boss informs you that you have to go to Ericsson Stockholm to give a

presentation You spend a few hours preparing the presentation, which remains on a

server on your home network You then take your lap top computer across the Atlantic to your meeting in Stockholm There you plug in to the LAN Without having to change any configuration parameters or addresses, you can immediately access your home LAN in Dallas and can easily give your presentation Also, external users can still contact you with your original IP address, as that has moved with you This transparent mobility is the job of Mobile IP

So, how does it work? Well, conceptually it's quite simple Here you can see your home network in Dallas and your foreign network in Stockholm When you are on your home network you are addressed using the IP address 193.234.210.74 Users on your own

network or on other networks connected to the Internet can talk to your computer using that

IP address, 193.234.210.74 The router on your home network ensures that packets bearing your IP address are delivered to you

Now, what happens when you pick up your laptop, spend 10 hours on a plane and arrive at Ericsson in Stockholm? Well, when you plug your laptop into this new foreign network your laptop talks to the router on the foreign network, which is called the Foreign Agent

Your laptop basically sends a message to the Foreign Agent saying, "Hi, I'm laptop

193.234.210.74 and I'd like to use your network." The Foreign Agent then sends a message

to the router on the home network, the Home Agent, and says "The mobile with IP

address 193.234.210.74 isn't on your network anymore, but it can be reached by sending packets to me at IP address 193.17.213.64." The Home Agent accepts this request and then we're ready to roll

So, what happens when some host out on the Internet starts to send packets to your

laptop? Well, the routers on the Internet have no knowledge of your boss's decision to send you to Stockholm, so they forward the packets to your home network as usual The router

on your home network, the Home Agent, sees these packets and remembers the care-of-address it received from the Foreign Agent So, it takes your IP packets and sends them through an IP tunnel over the Internet to the Foreign Agent The Foreign Agent looks at the packets and thinks, "This IP address isn't on my network, but I remember receiving

information about it, so I can forward it on directly." In this way packets reach the mobile node

What happens when the mobile node wishes to send packets out to the Internet? Well, this causes no problem whatsoever, the laptop builds IP packets with the desired destination address and the usual source address (that is, 193.234.210.74) and sends them directly out to the Internet as usual

Of course Mobile IP isn't perfect One of the biggest drawbacks is the routing inefficiency it introduces Packets destined for our laptop pass through the Internet to our home network and then back out onto the Internet to our foreign network This doglegged routing means that received packets suffer from higher latency than normal It also means that the Internet

is loaded more as the same packets must pass through it twice There are also some

important security implications to consider with Mobile IP The mobile node must

be carefully verified to ensure that packets are not forwarded to impostors intent on stealing others information

Mobile IP is only one of a host of macro mobility techniques Others, such as Dynamic DNS, L2TP and IP security are also capable of solving the problem of nodes moving

between interconnected networks

Ch 14 - Wireless and Mobile Data Characteristics

In most cases a mobile node wishes to talk to some remote computer For users and

applications, the way these computers communicate doesn't matter that much Using a wired

or wireless network makes no difference as long as the packets arrive This is why these industry standard protocols, such as TCP and IP, are used If the applications see some standard interface then they don't need to know if they are using a wireless network or a fixed network, they just see a standard interface These standard interfaces are extremely important to both application and network designers

It is extremely important to understand the difference between packet switched and circuit switched systems, a difference, which is illustrated here

You can compare a circuit switched data to a train and the physical medium is the train line Before the train (data) can use the line it has to ask and be granted access Once this is done the train can then proceed on the line, but it has exclusive access to the line No other trains can use the line while our train is there In a data communications environment,

circuit switchings main drawbacks are the call setup overhead (that is, asking for and being granted access to the line) and the low utilization of the physical medium (only one train at a time) One advantage of circuit switching is that the call setup procedure allows users to

be allocated and guaranteed a certain bandwidth A good example of circuit switching is the analogue telephone network Before you can talk to anyone you have to dial their number (a call request) The call is physically switched through the network until you had an end-to-end circuit between your telephone and the telephone you are calling Only when the phone

is answered at the other end (a call accept) can communication begin

What about packet switching? Well, in this case our data can be imagined as cars and the physical medium as a motorway

Instead of sending one big chunk of data, like we did with a circuit switched train, we send smaller chunks of data, with many users on the same channel simultaneously The advantage with this is that we don't need any call set-up procedure, we just jump in the car and go Packet switching also gives a much higher channel utilization, because many people can use the same motorway simultaneously

Ch 14 - Wireless and Mobile Data Characteristics

The disadvantage is that because we have no way of knowing how much data other users are going to send there is no guarantee of bandwidth if too many people try to use the

motorway at the same time it will slow down and eventually stop The best example of a packet switched network is the global Internet

There are lots of different wireless data communication systems To help us understand them all we divide them into different categories, Local Area Networks, Metropolitan Area Networks and Wide Area Networks Some of the techniques used for wireless LANs and MANs are shown here, but we'll discuss those in a little more detail in the next slide

Wide Area Networks have been divided up into two distinct groups here circuit switched and packet switched On the circuit switched side are Satellite phones, the American mobile telephone standard AMPS, GSM and PDC HSCSD is a new data oriented circuit switched service for GSM, which will be discussed a little more in chapter 16

The oldest of the packet switched services is Mobitex; a dedicated packet switched network that has been available since 1986

The other three systems are conceptually very similar, they all build on existing circuit switched mobile telephony infrastructures and they all add a packet switched service GPRS provides packet data services by building on the GSM infrastructure, PPDC provides packet data services by building on PDC and CDPD provides packet data services by building on the AMPS system

Ch 14 - Wireless and Mobile Data Characteristics

In the wireless LAN marketplace there are two main solutions, using infra-red or radio waves None of these solutions pretends to offer mobility - they merely offer a slightly more flexible way of coupling your desktop computer to the LAN In some cases this

wireless flexibility can be very desirable - laptop users do not need to continually plug and unplug network cables, they can simply move into the covered area and start working In some older buildings, pulling cable to every desktop can be expensive, so using a wireless solution can save money Of course, the usual drawbacks apply - the bit rate is much lower than that offered by other LAN technologies Typically IR and RF based LANs offer a bit rate of about 4 Mbps Compared with Ethernet, which is over 25 years old and yet manages

10 Mbps, that can be a considerable drawback IR based nets are also further limited by the fact that their connections must be "line-of-sight" based This is not a problem for RF based LANs

An interesting solution to the MAN question has been developed by an American company called Metricom They've developed a wireless system that breaks away from the usual wireless network topology that we saw in chapter 14.3 Instead of building a hierarchical system consisting of base stations, serving nodes and a backbone, Metricom build a semi-intelligent mesh consisting of small boxes that sit on top of street light and utility poles One

or more of these pole top boxes is equipped with a fixed connection to the global Internet When a user sends data it goes to the nearest pole top box This box then independently routes the information to a box that has a fixed connection All this routing is done

automatically by the meshed network, there is no human intervention and no need for

hand crafted routing tables