Riassunto Wireless Network, Appunti di Architettura Dei Sistemi Elettronici

Scarica Riassunto Wireless Network e più Appunti in PDF di Architettura Dei Sistemi Elettronici solo su Docsity!

Università degli Studi di Padova Dipartimento di

Matematica

Corso di Laurea Magistrale in Informatica

Wireless Networks

v0.1. A.A. 2017-

For new releases visit https://github.com/Polpetta/WirelessNetworkNotes/ releases

Contributions and issues reporting are welcome. Emanuele Carraro, Davide Polonio, Vassilikì Menarin, Eduard Bicego: Wireless Networks, c CC-BY-SA-4.0, 2018.

10.2 ZigBee Topology............................ List of Tables

1.1 Multimedia Requirements....................... 2 2.1 Wireless LAN Standards........................ 4 8.1 Comparison between different TCP version.............. 58 9.1 List of most important 802.11 wireless standards........... 60 vii

BER 10-3 10-6 10-6 10-

Data Rate 8-32Kbps 1-100Mbps 1-20Mbps 32-100Kbps Traffic Continuous Bursty Continuous Continuous Table 1.1: Multimedia Requirements

- Bit Error Rate: we assume that, when we lose a bit, we lose the whole packet.Voice packets are sent very frequently and are very small, that’s why we have a high BER, while for video the packets are bigger. Games are similar to VoIP. - Data Rate: indicates how much bandwidth we are consuming. - Type of Traffic: for voice, video and games, we have a device transmitting an uninterrupted stream (even though video only transmits in 1 direction). Data is different, we don-t care for every single packet we are downloading, we are interested in the total time of the complete thing. That’s also why we require 0 packet loss. Crosslayer Design When dealing with the different network layers, we can decide to keep them strictly separated, with no communication and information exchanged between them. That means, we have different layers with different functionalities that work together without overlapping. The Crosslayer Design blurs the separation between layers, allowing information in one layer to become available to other layers too. This means we adapt across deign layers, reducing uncertainty through scheduling and providing robustness via diversity. On one hand, one could argue that this messes up and complicates the code but, on the other, better communication can lead to higher performances. In WN, Crosslayer Design is not seen badly and therefore quite commonly used.

CHAPTER 2

Current Wireless Systems

We now present a brief review of some of the most commonly used wireless systems.

2.1 Cellular Systems

Cellular Systems are based on cells. Each cell has a different signal from the ones in its immediate vicinity to allow for seamless transitions (see the different colors in Figure 2.1). Frequencies, time slots and codes are reused at specially-separated locations. At

the center of each cell there is an antenna, certain antennas are more powerful than others and can manage hand-offs and control functions. These antennas are called Base Stations. Hand-offs can be can be horizontal if they occur between the same technology (4G → 4G) or vertical otherwise (3G ↔ 4G). Having larger cells means having lesser transitions but also more people accessing the same cell. If there are too many people and there isn’t enough bandwidth you won’t be able to use your device. Shrinking cell size increases capacity, as well as networking burden. Figure 2.1: Cellular Systems 3 4 CHAPTER 2. CURRENT WIRELESS SYSTEMS Figure 2.2: Wireless Local Area Networks (WLAN)

2.2 Wireless Local Area Networks (WLAN)

WLAN is used to connect local computers (within 100m range). We have an access point, connected to the Internet through a wire and using protocol 802.11. The access point not only forwards data, it also coordinates users and breaks data into packets. In fact, even though data arrives already broken, access points break it further, because the smaller the packet transmitted, the better. Obviously, this overhead becomes a problem when we relate it to overhead : since overhead has a fixed size, independent from the packet’s size, we may find ourselves in the situation where, faced with 1000 data +40 overhead we end up with 500 data +40 overhead splitting plus 500 data + 40 overhead. We still prefer splitting because, when facing a big error rate, we don’t want to lose all the packets. The WLAN channel access is

CHAPTER 3

Emerging Wireless Networks

We now further our system overview into some new, emerging systems.

3.1 Ad-hoc Networks

Ad-hoc wireless networks are an emerging wireless system, initially developed for military purposes. Ad-hoc networks are networks without a fixed infrastructure, Ad-hoc networks where you don’t have an access point, nor a backbone infrastructure to your network. The network is made up of equal nodes and based on peer-to-peer communications. To reach nodes that are too far away from the sender, we exploit multihopping. We could memorizing the network’s routing scheme, but note that we have a dynamic network topology! So, since you don’t usually know where you’re going, you can’t use the more common routing protocols. Ad-hoc Networks are very flexible, useful for many emerging applications. The capacity of such networks is generally unknown because depending on which node is transmitting, the nearby nodes will be blocked to avoid interference. Consider this structure (capital letters are nodes, arrows represent links): B ↔ A ↔ C ↔ D If D wishes to transmit to C while A transmits to B, it wouldn’t be able to do it because C is already occupied with A’s messages (interference!). Furthermore, we never really know the size and routing of our network, both can change very frequently. In this kind of networks crosslayer design is critical for efficiency crosslayer design and very challenging. Finally, energy constraint impose interesting design tradeoffs for communication and networking: for example, choosing to increase transmission power to avoid multihop leads to problems both in energy consumption (remember, if the distance doubles, the energy required is four time greater) and interference. I may occupy the whole network for just one transmission.

3.1.1 Sensor Networks

Sensor Networks are a specific type of ad-hoc networks. You have different sensors, each used to monitor a specific thing. The driving constraint is energy, since sensors 7 8 CHAPTER 3. EMERGING WIRELESS NETWORKS

3.2. MESH NETWORKS 9

Figure 3.2: Vehicular ad-hoc Networks (VANET) in the same direction move together in respect to each-other. The technology for automated vehicles is already present, the two main concerns are:

1. Legal problems: in case of accidents, whose fault is it?
2. Moral decisions: if a car has to steer to avoid an accident but in doing so it hits a passer-by, what is the best decision? Accidents are estimated to occur mainly when having automated cars mixed with regular cars (SW malfunctioning or hijack may still occur but are the secondary concern now). In a controlled environment we can monitor, while people are unpredictable. Finally, there is already a frequency reserved for vehicles communication (Standard IEEE 802,11p).

3.1.3 Opportunistic ad-hoc Networks

Opportunistic Nets were driven by commercial application needs and designed for when there may be an available Internet connection, but we try to replace it opportunistically with an ad-hoc network. Basically, they don’t use homogeneous nodes (like in traditional ad-hoc) but, instead, always try to use the best available. Their main feature is flexibility.

3.2 Mesh Networks

Mesh Networks are something between ad-hoc networks and wireless ones. They are different because the can connect to both regular and completely wireless access points. In this sense, they are ad-hoc opportunistic extensions of a fixed urban infrastructure. Their main purpose is to create a low-cost, easily deployable, high performance wireless coverage. When designing a Mesh Network, one must carefully choose the routing protocol, considering all the types of connection you may need. Routing protocols have to achieve fairness and local balancing, also remembering that in multihop, having two nodes communicate may prevent others in range. Finally, it is difficult to determine a quality of service, that is stating a quality standard and being able to assure it.

10 CHAPTER 3. EMERGING WIRELESS NETWORKS

Figure 3.3: Mesh Network (the green ones are regular access points)

3.3 Distributed Control over Wireless Links

These networks deal with automated vehicles: cars, Unmanned Airborne Vehicles (UAV), insect fliers and so on. In these kind of networks it is important to have a good control of the environment, so it is important to avoid packet loss and/or delays and also assuring a good bandwidth and speed when delivering messages. Avoid queues since they cause delays. The controller design should be especially robust to network faults.

3.4 Radio Frequency Identification (RFID)

RFIDS are based on magnetic fields and are usually tags without a battery (even though they can have it). They store more info than a simple barcode and are therefore used in supply chains instead of them. They don’t need a direct optical reading but, instead, require an emitter of electro-magnetic waves that charge the tag. The charged tag sends a message containing all the information back to the server, that can then check it. Even though it contains more information that a simple barcode, it is more expensive. Their popularity is increasing, and standards are currently under development.

3.5 Nano-Networks

Networks made up of very tiny objects and very small distances. You can’t just replicate in scale what you would normally do but instead you have to rely on different means of communication, very popular are calcium signals. 3.5. NANO-NETWORKS 11

CHAPTER 4

Physical Layer

4.1 Radio Signals

Most of the wireless technology used today is based on radio frequency signals. Radio signals are electromagnetic energy generated by high frequency alternate current in antennas. They can have different frequencies and lengths and propagate differently depending on the medium they need to cross. Radio frequencies have three main properties:

- Amplitude : determines how far a signal can travel (remember that reaching 2x distance requires 4x power). It is the difference between the highest and lowest wave peak. - Frequency : the number of oscillations in 1 second - Wavelength : the distance between two high (or low) peaks. The wavelength is (^) frequency c (^) and this means that, in practice, antennas work better with sizes 1, , of wavelength. Propagation From its origin, after a certain point the signal is no longer detectable. If there is an obstacle the signal loses power more quickly because obstacles can reflect or absorb waves depending on materials and waves frequencies. When dealing with RF propagation we can rely on a rule of thumb: high frequencies are good for short distances and are affected by obstacles, while low frequencies are good for long distances and are less affected by obstacles. RF have other properties that describe their behavior: they can phase, that is done shifting the wave (in degrees or radians) and are affected by the physical orientation of the antenna (polarization). Waves propagate in a toroid form (donut form). The propagation range depends on power, obstacles, the receiver’s sensitivity and many other factors. We have different ranges: - Transmission Range : marks how far can the communication reach; - Detection Range : marks how far the detection of the signal is possible; 13 - Interference Range : marks the distance at which the signal is too far away from the sender to be detected and so it just adds to the background noise. physical effects Also, a lot of different physical effects can affect propagation, in particular: shadowing, reflection, refraction (happens when passing to a different medium), scattering (at a