A network is a group of interconnected Network Elements. The interconnections among the Network Elements can be implemented by: • Optical fiber • Microwave
The N Es can be connected with different architectures: • tree • ring • mesh Normally the network has a • hierarchical architecture structured in layers The N Es in the lowest layer are connected each other and one N E (or more than one), operating as Hub, gives the connection to the higher layer. In the next page an example of hierarchical network is shown.
Layer 0 is the lowest layer implemented in the Local area. The typical architecture of Layer 0 is Tree. Normally the interconnections in this layer are made with microwave links. Layer 1 is the intermediate layer implemented in the Regional area. The typical architecture of Layer 1 is Ring. The interconnections in this layer can be made by optical or microwave links. Layer 2 is the intermediate layer implemented in the National area. The typical architecture of Layer 2 is Mesh. The interconnections in this layer normally are made by optical links.
In the Mobile Networks in the Local area there are the Node B, in the Regional area the Radio Network Controllers or Mobile Management Entities and in the National Area the Core Network with the Mobile Switching Centers. Depending on the capacity to increase the resilience between a lower layer and an upper layer it is possible to have one or two links, as shown in the figure.
In Step 1 we start with the Network with all the stations. In Step 2 we draw all the possible hops interconnecting one station to all the other stations. This step is done during the site survey: from one station it is checked, if there is the optical visibility between the stations. The length of each possible hop is also calculated. In Step 3 by hand or by using a software algorithm, if the network is huge, the best tree interconnecting all the stations is calculated. The best tree is the tree with the minimum cost: normally the cost is the length of the hop.
The problem is how to connect all the Network Elements with the minimum total length of hops and decide which hop must be used to connect all the Nodes with the minimum total length. If the network consists of few Network Elements, the Tree can be built by hand, but if the network is made up of several Network Elements, it is necessary to use an algorithm to built up the Tree with the minimum cost. The used algorithm is the Kruskal algorithm. In Internet you can find a huge literature about this algorithm. Select the hops in the Network one by one, starting from the hop with minimum length. Do no select an hop, when it causes a loop. The obtained Network is called Minimum Spanning Tree (M S P).
A Network with Nodes and Hops and the relevant costs between two adjacent nodes is given. We will apply the Kruskal algorithm. The Kruskal Algorithm belongs to the class of algorithms known as “Minimum Genetic Tree”. The algorithm starts with a Network, which consists of “n” Nodes. The Network is considered like a Forest and the Nodes are consider like the Trees in the forest. Initially, each Node is a tree, then a tree is formed by a set of Nodes connected to each other. At every step of the algorithm, two different Trees of the Forest are connected to a bigger tree. Therefore, we keep having less and bigger trees in our Forest, until we end up in a tree, which is the Minimum Genetic Tree. In every step we choose the side with the lowest cost (this policy is called “greedy policy”). If the chosen side connects Nodes which belong to the same tree, the side is rejected, and not examined again, because it could produce a loop, which will destroy our tree.
In the first step the lowest cost is 5 between Nodes J and I. In the second step the lowest cost is 5 from Nodes J and D. In the third step now the lowest cost is 10 between Nodes E and F. In the fourth step the lowest cost is 10 between Nodes C and K. We have to go on to connect all the nodes.
The final solution is obtained by continuing to select one hop at a time, always with the minimum cost, until all the Nodes are connected to each other.
Now if we consider the network, we find two crossed hops: AF and BG. Normally we have to avoid this situation, because a propagation problem, due to the rain, can affect at the same time both hops. Finally, we have to consider the adjacent hops and the parallel hops in order to assign correctly the frequency of the radio channels in order to reduce the interferences. This topic will be explained in the Radio Network Planning (Basic course) RN00021.
Now we have to select the Hub node, which will be connected to the upper layer. The candidate Hub node must have the requisite that all the Nodes must be connected to the Hub node with a maximum of 3 consecutive hops. This rule is not compulsory, but it is a general rule normally followed. In our example the candidate for Hub node is Node B. All the Nodes are connected to Node B with a maximum of 3 hops, except Node H, connected with 4 hops.
If the network consists of few N Es, the ring can be built by hand, but if the network is made up of several N Es, it is necessary to use an algorithm to built up the Ring with the minimum cost. The used algorithm is the Travelling Salesman. In Internet you can find a huge literature about this algorithm. The problem of building a Ring with the minimum length (or more generally with minimum cost) is one of the most known problems, named “Travelling Salesman Problem”. A salesman must visit several cities, passing through each city only once, beginning from one of them which is considered as his base, and returning to it. The cost of the transportation among the cities (whichever combination possible) is given. The program of the journey is requested, that is the order of visiting the cities in such a way that the cost is minimum. In our example, given in the next pages, the cities are replaced by the Network Elements and the cost is given by the length of the radio hop.
The Travelling Salesman algorithm is an heuristic algorithm which finds solutions among all possible ones but it does not guarantee that the best will be found, even if usually it finds a solution close to the best one. We start to draw the Line of Sight matrix (L O S). The matrix is a table, where for each N E are written all the visible N Es with the relevant hop length in kilometers. As example N E 1 can be connected to N E 2 with length 5 km, to N E 3 with length 3 km and N E 4 with length 2 km. The L O S matrix is filled during the site survey.
Now we go to the row of N E 3. The lowest cost in the row of N E 3 is 2 associated with N E 5. The third connection is from N E 3 to N E 5. Also N E 2 has the same cost, but if we select N E 2 and go on in the procedure, we can see that at the end we are not able to create a Ring.
We are now in the last step in the row of N E 2. The lowest cost in the row of N E 2 is 5 associated with N E 1. The last connection is from N E 2 to N E 1 and we have completed the Ring.
The radio hop must be designed in Line Of Sight. This means that there is electromagnetic visibility between the 2 radio stations. The electromagnetic wave propagation of the RADIO systems is in the lower part of atmosphere, near the ground. The presence of the atmosphere and of the ground can affect the R F propagation. The PROPAGATION depends on: • CLIMATIC CONDITIONS • R F FREQUENCY BAND • RADIO HOP LENGTH • GROUND CHARACTERISTICS
The Main Propagation phenomena depend on the Atmosphere, Rain and Ground. The effects of the Atmosphere are: • Atmospheric Absorption • Refraction through the atmosphere: Ray Curvature • Refraction through the atmosphere: Multipath Propagation. The effects of the Rain are: • Raindrop Absorption • Raindrop Scattering • R F Signal Depolarization. The effects of the Ground are: • Diffraction through Obstacles • Reflections
The nominal Rx power level is the Rx level signal, that we normally receive without any propagation impairment. During propagation the rain and the multipath activity can cause an attenuation of the Rx signal: this attenuation is called Fading. In some cases the multipath activities can cause not an attenuation, but an increase of the Rx signal: this increase is called Up-fading. When the Rx signal, due to the fading, goes below the Rx threshold, the radio hop is out of service: the relevant time is called Outage. The Rx threshold depends on the R F band and on the Modulation scheme and it is written in the Technical Specifications of the radio equipment. The difference between the Rx nominal signal and the Rx threshold is called Fading Margin.
To design a radio hop we have to consider the Outage periods. If the period is ≥ 10 seconds, this period is called Unavailability period. If the period is < 10 seconds, this period is called Quality period. In this case the hop is available, but with a not good quality.
Depending on the R F band used, we have different propagation impairments that affect the radio system. The impairments are due to the rain and to the multipath activity. As far as the propagation characteristics are concerned, rain attenuation predominates at frequencies >17 GHz, while multipath predominates at frequencies < 11 GHz. The radio system should be mainly designed in terms of Unavailability at frequencies above 17 GHz and in terms of Quality at frequencies below about 11 GHz, while in the range 11 to 17 GHz both objectives should be considered.
The impairment due to the rain depends on the wavelength of the radio signal and the size of the Rain drop. If the wavelength is bigger than the size of the rain drop, the impairment is small. If the wavelength is comparable with the size of the rain drop, the impairment is strong. This is the reason why the impairment is negligible for RF bands lower than 11 GHz and the impairment is strong for RF bands higher than 17 GHz.
When the raindrop fails, the shape changes from spherical to ellipsoidal due to the air friction. The horizontal size is bigger than the vertical size and also the impairment is bigger. For this reason it is better to use the vertical polarization.
The multipath activity is due to the stratification of the atmosphere. The radio rays transmitted can be reflected and in reception side the antenna receives the direct ray, but can receive also some reflected rays. Depending on the different phase and delay of the direct and reflected rays, we have an impairment in the total received signal. In a short hop, as shown in case 1, the probability that the reflected rays enter the Rx antenna is very low and consequently the impairment is low or null. In a long hop, as shown in case 2, the probability that the reflected rays enter the Rx antenna is high and consequently the impairment is high.
The Fading is an attenuation of the Rx nominal power due to the propagation impairments. There are two types of Fading: Flat Fading and Selective Fading. The Rain causes Flat Fading. The Multipath activities cause Flat Fading and Selective Fading.
With Flat Fading each frequency of the R F channel spectrum is attenuated by the same value. With a 60 dB Flat Fading each frequency is attenuated by 60 dB and no distortion in the shape of the spectrum is produced.
The Flat Fading is counteracted by the Low Noise Amplifier in the Receiver. The L N A has a variable gain in order to produce at the output always a constant power, suitable for the correct operation of the Demodulator.
The stratification in the atmosphere causes the multipath activity, which creates a distortion in the spectrum responsible of the Inter-symbol interference and at the output of the Demodulator we have errors.
The measurement of the Signature is made in the lab by using the Signature Measurement Instrument. This instrument sends to the input of the Demodulator a distorted spectrum with a specific frequency strongly attenuated (this is called Notch). The frequency of the notch is attenuated until at the output of the Demodulator a 10-6 B E R is generated and this attenuation value is a point of the Signature.
The Signature is calculated frequency per frequency of the spectrum. In Step 1 is attenuated the central frequency f0 until B E R 10-6 is generated. The Notch depth is the measured attenuation and this value is the first point of the signature curve. In the figure is 10 dB. In Step 2 is attenuated another frequency (f1) until B E R 10-6 is generated and this is the second point of the signature curve. In the figure is 5 dB. In Step 3 is attenuated another frequency (f2) until B E R 10-6 is generated and this is the third point of the signature curve. In the figure is 5 dB. And so on for each frequency.
At the end when the notch has been generated at each frequency of the signal spectrum, the signature can be drawn. The Minimum phase and Non-minimum phase signatures are equal in the modern demodulator.
The picture compares the signature of two demodulators produced by two different manufacturers. Signature 2 is better, because the Demodulator is capable to counteract a stronger attenuation due to the selective fading. If we consider the central frequency f0, in Signature 2 a B E R 10-6 is generated with a notch depth of 10 dB and not with a notch depth of 5 dB, as in Signature 1.
Final consideration: to know if the Demodulator is capable to counteract the Selective Fading in a good way, it is important the area of the Signature: smaller is the area and better is the Demodulator.
In the picture is shown an example of the signature of a Demodulator. The area of the Signature is calculated by using these two values: Bc and Delta f0
The ITU-R targets depend on: • the hop length and • the position of the hop in the radio network, depending if the hop is on the Local layer, Regional layer, National layer or International layer.
To counteract the Unavailability causes, we have to implement: • the battery protection for the failure due to the power outage • the card protection for the failure due to the radio equipment • a good fading margin calculation to eliminate or at least reduce the unavailability due to the rain. In this planning course we will consider only the problems due to the rain.
The Rx threshold depends on the Modulation scheme. The higher the Modulation scheme, the higher the relevant threshold. As it is shown in the figure, by using a lower modulation scheme (512 Q A M or 16 Q A M) the Unavailability period is reduced. This is the advantage, but the disadvantage is the reduction of the net throughput capacity.
The Quality is mainly due to the Multipath activity, which is caused by the stratification in the atmosphere. In this case the Receiver can receive not only the Direct ray, but also several Reflected rays. Depending on the phase shift and the delay between the rays the total received signal can be strongly attenuated and distorted. The stratification in the atmosphere is not constant, but it changes in a very fast way. For this reason this impairment does not cause the Unavailability, but only affect the Quality: this means that the Rx level is below the Rx threshold for less than 10 seconds.
How to counteract the Multipath activities? • Of course it is not possible to eliminate or reduce the multipath activity BUT • We have to consider the radio equipment capability to support the impairments • We have to consider the Adaptive Time Domain Equalizer (A T D E) inside the Demodulator, which performances depend on the signature of the Demodulator • If the A T D E is not strong enough, we have to implement the Diversity configuration
• The I T U Recommendations are only Recommendations. This means that it is not compulsory to follow them • The I T U Recommendations are defined on connection, where a connection can be formed by 1 or more radio hops. • For this reason, normally the Customer prefers to assign the same target to all the hops, regardless the length of the hop and the used R F band
The table shows some of the commonly used percentage values for unavailability. The five nines requirements, correctly or incorrectly, has been applied to almost everything.
The Unavailability target is given in minutes per year, because the unavailability is given by the Rain and for this reason we have to consider the Rain intensity in each month of the year. The Quality target is given for historical reason in seconds per month and must be consider the worst month of the year. Of course more stringent is the target, higher is the cost of the radio hop.
The Reliability is not an ITU approach. • The ITU performances are evaluated in terms of quality and unavailability • A different availability criterion is sometimes adopted in terms of annual percentage (99.xxx %): it consists of a weighted sum of ITU quality and unavailability. • This quantity, deriving from Bellcore standard, has been defined as Reliability in order to avoid confusion with the previous ones.
Two courses have been developed: • Radio Network Planning (Basic level) course (RN00021-K) • Radio Network Planning (Advanced level) course (RN00022-K) The Advanced level course includes an Appendix section (available in PDF format only) explaining in deep some topics. The Appendix section is divided in Appendix A and Appendix B.
1. Appendix A explains the following topics: • Basic notions: 802.1p vs. DiffServ • Microwave Packet Radio QoS implementation • Capacity planning 2. Appendix B explains the following topics: • Effects of the atmosphere • Theoretical Rx threshold calculation • Dispersive Fading Margin / Interference Fading Margin / Composite Fading Margin • ITU outage targets • Adaptive Modulation and Radio LAG in Network Design • Dual band antennas • Near field/Far field and Passive repeaters • Some notes on the Reflection from ground • Radio Network Planning summary • milliWatt and dBm and Acronym list
The block diagram shows the procedure to be followed in the Basic level planning course. In Step 1 the R F frequency of the channel will be selected. In Step 2 will be calculated the PRx nominal level of the signal by using the Power budget formula. The inputs are the Tx power, the antenna gain, the propagation losses and other losses. The diffraction of the electromagnetic wave due to the obstacle will be also considered. The output of step 2 is the Fading Margin. In Step 3 will be calculated the fading predictions due to the Rain and to the Multipaths. In Step 4 will be calculated the Quality and Unavailability outage targets of the radio hop. In Step 5 we will compare the calculated outage targets with our planned targets. If the outage targets are meet, we go to the next page. If the Unavailability outage target has not been meet, we have normally to increase the Fading Margin and go back to Step 3. If the Quality outage target has not been meet, normally we have to implement the Diversity configuration.
The block diagram shows the procedure to be followed in the Advanced level planning course. In Step 6 we have to consider the Frequency re-use configuration, if this configuration is used. In Step 7 the interferences created by our radio network will be considered. In Step 8 will be calculated the degradation in the Rx threshold, due to the interferences. In Step 9 if this degradation is acceptable, the planning is finish, otherwise we have to solve this problem; normally by changing the polarization or the frequency of the R F channel.