Network Performance - Science topic

Dear Sami Naili

Your question is interesting. You may go through the following articles for your solution -

It is better you use industry standard tools like MATLAB or NetSim.

You can look at this link https://support.tetcos.com/support/solutions/folders/14000128567 and NetSim documentation for 5G at https://tetcos.com/help/v13.2/Technology-Libraries/5G.html

Hello Dr. James Marcus Griffin

Thanks for the question and the clarification answer.

Best regards

One possible reason could be that many smaller files involve that mean TCP connection establishment and tear downs, while one largefile would have only one TCP establishment and tear down. So the overhead when having many small files is larger.

Amro Saleh there is no need to convert a gray scale image to RGB, it will definitely increase your computation cost as Ken Mascola says. You can directly pass a grayscale image to CNN by setting your channel dimension as 1. Check this link:

https://yashk2810.github.io/Applying-Convolutional-Neural-Network-on-the-MNIST-dataset/

Your problem is as simple as working in a MNIST dataset that has only grayscale images.

I would go with MATLAB for 5G link level simulation, or NetSim if you want to do 5G system level simulation

In order to measure the distance between two nodes communicating with each other in full duplex mode then as Manfridi said you can send a sequence of bits and receive them again promptly from the the other node without and processing them.

Then the transmit and receive time will be delay time between the front of the transmit signal and the front of the received signal. This time = 2L/v whwere L is the length and v is the velocity of the transmission. If the signal is propagating in space its speed will be = speed of light in the space c. In other medium the v= c/sq.root of epsilon r where epsilonr is the dielectric consatnt.

Best wishes

Dear Fawaz,

One method to access the available wireless networks are supplying the mobile devices with the capability to detect the different networks and display them for the users and the user can choose one from them. Most advanced mobile user equipment and laptops are supplied with this property. They detect the Wlans and they detect the the mobile service if it contains a valid simcard.

As a user you can choose among them.

As for the mobile devices there is a need for the hand off control between the wide area network and the wlan. This type of handover is called vertical handover and there are some protocols developed for it in the so called heterogeneous network.

Best wishes

Emanuele Dalsasso fair enough!

How can make security settings for APS in opnet project

Dear Yashuai,

welcome

In addition to the construction of the neural network, i think that the error can be due to nonlinear behavior of the inverse tan function as becomes very nonlinear after 45 degrees. In additions the values of the tan becomes large. For Cartesian to polar conversion you can use Cordic algorithm which is considered one of the efficient algorithms for the conversion process.

Best wishes

Any related work on Throughput utilization of WSN for worst case situations?

Dear Hantouti,

in the following literature You find the required informations.

Here are all link-based and node-based metrics presented:

 R. Baumann, S. Heimlicher, M. Strasser and A. Weibel: A Survey on Routing Metrics, TIK Rep. 262, Computer Engineering and Networks Laboratory, ETH-Zentrum, Switzerland, Feb. 2007

 http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.304.7863

 G. Parissidis, M. Karaliopoulos, R. Baumann, T. Spyropoulos and B. Plattner: Routing metrics for Wireless Mesh Networks, in Guide to Wireless Mesh Networks, Springer, 2009

 http://cgi.di.uoa.gr/~mkaralio/wp-content/uploads/papers/B1-Routing_metrics_for_Wireless_Mesh_Networks-Springer08.pdf

L. Zhao and A. Y. Al-Dubai: „Routing Metrics for Wireless Mesh Networks: A Survey“, Lecture Notes in Electrical Engineering (LNEE), Part of the Lecture Notes in Electrical Engineering, Vol. 127, Feb. 2012

 http://www.napier.ac.uk/~/media/worktribe/output-293859/csieliangzhaopdf.pdf

N. Baccour, A. Koubâa, L. Mottola, M. A. Zúñiga, H. Youssef, C. A. Boano and M. Alves: „Radio Link Quality Estimation in Wireless Sensor Networks: A Survey“, ACM Transactions on Sensor Networks (TOSN), Vol. 8(4), Sep. 2012; DOI: 10.1145/2240116.2240123

 http://dl.acm.org/citation.cfm?id=2240123

For the information on ETX see:

 D. S. De Couto, D. Aguayo, J. Bicket and R. Morris: A High-Throughput Path Metric for Multi-Hop Wireless Routing,  MobiCom ’03, Sep. 2003

 https://pdos.csail.mit.edu/papers/grid:mobicom03/paper.pdf

N. Javaid, A. Javaid, I. A. Khan and K. Djouani: Performance Study of ETX based Wireless Routing Metrics, 2nd IEEE International Conference on Computer, Control and Communication, Feb. 2009; DOI:  10.1109/IC4.2009.4909163

 http://ieeexplore.ieee.org/document/4909163/

 For a well-founded mathematical description of metrics see:

Y. Yang and  J. Wang: Design Guidelines for Routing Metrics in Multihop Wireless Networks, IEEE 27th Conference on Computer Communications, (INFOCOM 2008), Apr. 2008; DOI: 10.1109/INFOCOM.2008.222

 http://ieeexplore.ieee.org/document/4509817/

P. Karkazis, P. Trakadas, H. C. Leligou, L. Sarakis, I. Papaefstathiou and T. Zahariadis: Evaluating routing metric composition approaches for QoS differentiation in low power and lossy networks , Wireless Networks, Vol. 19(6), Aug. 2013; DOI 10.1007/s11276-012-0532-2

 https://link.springer.com/content/pdf/10.1007/s11276-012-0532-2.pdf

Best regards and much success

Anatol Badach

Dear Rahul,

The colleagues above gave a satisfactory answer to your question but i want to stress some basic principles: 

Jitter concerns statistical variations of theoretically regular occurrence of certain event.

For example the period time of an oscillator. If the oscillator is ideal then its pulses will be very strictly eqi spaced in time, however, this will never physically occurs because there is statistical factors affecting the periodic time such as the noise signal which varies the zero crossings of the deterministic waveform.

Like wise the time jitter  of the average delivery time of the packets in a packet switched networks. Because the packets transport in packet switched network has statistical nature therefore the packet delivery will not be strictly regular. The delay  gitter may be defined by the standard deviation of the delay from an average value.

It is calculated as it is defined.

best wishes

To answer completely to your question more detail on the measurement are necessary, and first of all the real observed differences in terms of relative ratios. In my opinion variations of few percentages are acceptable for “field measurements in exact site”. Anyway, I send you a general response to give response to yours question.

In physical sense, the seismic measurement is the result of a chain of many transfer functions (TF). Namely, using their frequency (f) domain representations:

source S(f), regional propagation model M(f), local site propagation model L(f) and apparatus A(f). A(f) can be further splitted in the sensor TF, I(f) and data-logger TF, DL(f). so that our measurement in frequency domain is the product of them (or convolution in time domain).

When we perform a seismic measurement we must take into account the contribution of each component (TFs).

Now we consider the measurement with “two seismic sensors, of different brand but same bandwidth, in the exact same location”. In this case we can assume that S(f) (the transfer function of a set of sources, (which are homogeneously distributed around our measurement point) and M(f) are the same. In other words, the input at the measurement site can be assumed the same for the two sensors. If we install the sensor at “exact same location” (==really a small distance between them) with the same terrain-coupling procedure, we can assume that L(f) is again the same for the two sensor. Then the measurement should depend on only the apparatus transfer function A(f)=I(f)*DL(f).

If the cited conditions are satisfied, the answer to your question could be find in the responses of sensor and/or data logger. In particular, sensor response I(f) mainly depends on: the type of sensor (velocimeter, accelerometer) and its constructive modality (passive or active: generally, force balanced type). DL(f) is mainly depends on the Analogic/Digital conversion system, which involves the frequency response of: cable and connectors, the embedded electronic of amplifiers of data-logger system, the sampling rate and decimation procedure.

Assuming that the sensors are of the same type and have the same constructive modality we should invoke the effect of data-logger response. Assuming the same sampling rate of A/D converter and decimation procedure (you should verify this condition, also due to different producers of stations) the main effect should be produced by the data-logger (embedded) electronic that, generally, it manifests as the presence of continuous component in recording (also variable in time). This effect is relatively relevant in presence of low noise site recording.

If the data-logger and cable are the same we must invoke the effect of sensor response. You must verify if the manufacturer responses in your bandwidth are the same. Finally, an operational consideration: using sensors with different proper period it is necessary to wait to reach the optimal recording condition. Also in this case spurious response, like variable continuous component and mainly in low frequency, could be observed.  It is important, to perform high quality seismic noise measurements, to use a very long recording windows (i.e. almost 1-2 days).

Dear Choi,

You can make a survey on previous research articles related to S-ALOHA and review CDM-based results and discussion. 

Well i don't know how to start Mr. Shilpy Agrawal. Offloading is a kind of handover, but based on QoS and not for example on SINR. So when an eNB is overloaded, some of the users have to be handed over to another eNB or a Femto and so on, and this to release some of the radio resources to maintain the QoS of the rest of the users. So really, its not  looked on as the max value of throughput of the eNB. But on how much resources the eNB can assign to the users, and this depends on the BW and the number of users. So to start with, just find the number of resource blocks (RB), and you can give a random  QoS requirement for each user (number of RBs) and then see if all users are satisfied, other wise the eNB is congested or overloaded. In this case handover  one user and calculate satisfaction rate, is it still below? yes , handover another user, if No stop handing over. That's it. very simple. Hope this helps.

//** Apologies, if you have received multiple copies. Please forward the email to interested colleagues and research scholars**//

Respected Sir/Madam,

Book Chapter proposals are invited for the Edited book titled "Opportunistic Networks: Mobility Models, Protocols, Security & Privacy" in the following topics:

1. Foundations of Opportunistic Network:

v  Migration from MANET to Opportunistic Networks, Basic Concepts of Opportunistic Networks (OppNets)

v  Need of Opportunistic Networks, Types of Opportunistic Networks

v  Sensor Networks: MULE, BIONETS, ZebraNet, SWIM 

v  Pocket Switched Networks: MANET, MSN, VANET, DTN

v  Amorphous Opportunistic Networks: Semantic Opportunistic Networks, Freenet

v  Requirement for OppNets

2. Delay Tolerant Networks (DTNs):

v  Basic Concept, DTN Architecture: Virtual Message Switching Using Store-and-Forward Operation, Nodes and Endpoints, Endpoint Identifiers (EIDs) and Registrations, Priority Classes, Congestion and Flow Control at Bundle Layer

v  Buffer Management, Challenges for Applications over DTN

v  Security for DTN Applications

v  Case Studies: Message-Based Applications—E-mail/SMS, Stream-Based Applications—XMPP, Web, Content Search, Underground Mines, Floating Content

3. Taxonomy of Mobility Models:

v  Random Waypoint Movement Model (RWPM)

v  Map-Based Movement Models: Well Known Text (WKT) format: Map Route Movement and Shortest Path Map Based Movement (SPMBM)

v  Route Based Movement Model (RBM), Working Day Movement Model (WDMM)

v  Transportation Analysis and Simulation Systems (TRANSIMS)

v  Social Network Based Mobility Models: Social Network Models, Community Based Mobility Model

v  Pursue Model, Manhattan Mobility Model, Rush Hour (Human) Traffic Model, Random Direction Model

v  Brownian model, Obstacle Mobility model, Trace-Based Mobility Models, Smooth Random Mobility Model

v  Evaluation of Mobility Models

4. Taxonomy of Routing Protocols:

v  Infrastructure Opportunistic Networks

v  Dissemination based: Epidemic Routing: PROPHET, MaxProp, MV Routing, Spray &Wait, SMART, MARS, RAPID, EBR, SCAR, Direct Delivery

v  Network Coding Based Routing

v  Context Base Routing: Context –Aware Routing (CAR) Protocol, PROPICMAN, Moby-Space Routing Protocol, History-Based Opportunistic Routing Protocol (Hi-BOP)

v  Without Infrastructure Opportunistic Networks

v  Evaluation of Routing Protocols

5. Route Optimization Techniques for fast Data Delivery:

v  Basic Concepts, Route Optimization using Membrane Computing, Route Optimization using Game Theory, Route Optimization using Ant Colony Optimization (ACO), Route Optimization using PlayNCool

v  Performance Evaluation of Route Optimization Techniques

6. Improve Quality of Service of Opportunistic Network:

v  Heterogeneous Connections: Connection Process, Heterogeneous Connections, Network Capacity

v  Network Selection: Network Discovery, Network Selection Algorithms for Information Dissemination

v  Message Forwarding: Single-Copy Forwarding, Multi-Copy Forwarding

v  Congestion Control, Mobile Agents   

7. Vehicular Ad hoc Networks (VANETs):

v  Basic Concepts, VANET Architecture

v  VANET Routing Approaches: Geocast/Broadcast, Unicast, Multicast

v  Geocast/Broadcast:Spatially aware packet routing algorithm, SADV, Interference Aware Routing, FROV,Multihop broadcast protocol, Proposed Algorithms: V-TRADE, UMB, AMB, MHVB, MDDV

v  Unicast:  Greedy: GSR, GPCR, GyTAR, Connectivity Aware Routing, Opportunistic: OPERA, Topology-assist geo-opportunistic routingandMaxProp, Trajectory: SiFT, GeOps, TBD, TTBR

v  Multicast: Topology-Based Approaches: ODMRP, MAODV, GHM, Location-Based Approaches: PBM, SPBM, LBM, RBM, IVG

v  Simulation Methods: Traffic Simulators,Network Simulators

v  Applications

8. Energy Management:

v  Localized Algorithms save Energy, Minimum-Energy Broadcasting and Multicasting, Power-Aware Routing

v  Power-Efficient Neighbour Communication and Discovery for Asymmetric and Symmetric Links

v  Challenges for Energy Management

9. Network Coding Schemes:

v  Basic Concepts

v  CodeOR, SlideOR, AONC, BEND, COPE

v   Performance Evaluation of Network Coding Schemes

10. Taxonomy of Security Attacks

v   Basic Concepts of Attacks, Types of Attacks: Jamming: DoS, DDoS, RDDoS Attack, Black Hole Attack, Cooperative Black Hole Attack, Wormhole Attack, Byzantine Attack, Sybil Attack, Sleep Deprivation,  Rushing Attack, Colluding misrerlay attack, Link Spoofing Attack, Selective Forwarding Attack (Gray Hole attack), Snare Attack, Blackmail Attack, Cooperative Blackmail Attack, Cloning Attack, Sinkhole Attack, Desynchronization attack, Overwhelm attack, Jellyfish Attack

v  Comparison of Attacks

11. Security, Privacy & Trust, Anonymity:

v  Security issues, Authentication and Authorization, Privacy and Trust management techniques, Identity management,  Location-based security services

v  Mix Networks, Intrusion Detection and Prevention Techniques

12. Future Networks Inspired by Opportunistic Networks:

v  Introduction

v  BIOlogically-inspired autonomic NETworks and Services (BIONETS),Mobile Ubiquitous LAN Extensions (MLUEs), Shared Wireless Infostation Model (SWIM), Vehicular Delay Tolerant Networks (VDTNs), ZebraNet, Freenet,

DakNet

v  Evaluation of OppNets

13. Network Emulation Testbed for Opportunistic Networks:

v  Design Challenges: Mobility, Disruptions, Wireless Communication, Scalability,

v  QOMB Testbed: StarBED, QOMET, Overall Architecture

v  DTN Implementations

14: Applications and Future Trends:

v  Introduction

v  Opportunistic Computing, Mobile Data Offloading, Research Trends on Recommender Systems

15: Hands-on “The ONE simulator” – Opportunistic Network Environment

Important Dates:

Last Date for Submission of Chapter Proposal (Minimum 4 Pages)                01 August 2017

Acceptance/Rejection of Book Chapter Proposal notification to Authors         20 August 2017

Full Chapter submission by Accepted Authors                                      10 December 2017

Submission of revised/Final Chapters to publisher                                 10 January 2017

NOTE: The book chapter proposal may kindly be sent by strictly following the attached format to the following Email Id: khaleelahmad.csit@gov.in

EDITORS:

Dr. Nur Izura Udzir

Universiti Putra Malaysia, Malaysia (nurizura@gmail.com, izura@upm.edu.my)

Dr. Khaleel Ahmad

Maulana Azad National Urdu University, India (Khaleelamna@gmail.com)

Dr. G C Deka

Directorate General of Training, Ministry of Skill Development & Entrepreneurship, New Delhi-110001, India (ganeshdeka2000@gmail.com)

The standard data rate of 1 Mbps and ( cough! ) "enhanced" data rate of 2Mbps is pretty much what you have to work with.  The high speed devices are still fairly expensive and uncommon, as you mention, those get you to 25Mbps.    That is the native symbol rate.  So not much to "fix" there.

As for Bluetooth cutting out?  That's not a feature of Bluetooth, it could be a number of factors specific to your use case: drivers, RF environment, the transport methods, the computing devices.  I consider that not to be an intrinsic Bluetooth issue.   For synchronous transfer near maximum throughput, you do not have an throughput to retry lost block because the aggregate of valid and retried data exceeds the available throughput.   So you cannot run at 1Mbps, as there is no room for errors.   The strategy is to reduce the bulk of the data, leaving ample headroom.

For asynchronous data transmission, you need s TCP-like transport mechanism.  Heck, even things like Z-Modem does a block integrity check and re-transmission, is appropriate.

Another factor is the RF space.   Bluetooth devices come in 3 distance ranges - in some cases the only difference is the type of antenna in the device.   The maximum range of 100 meters is achievable.   Its also possible to attach yagi antennas to bluetooth interfaces dramatically extend the range.   I have, myself, sent bluetooth over 500 meters with very cheap yagis.  You can even do this with low power devices.  More gain is always better,

The point of the above paragraph is that if you improve the gain of the antennas, you can have much more immunity to ambient RF.   Improving the SNR will help increase the throughout.

Hope that helps.

NS2, NS3 & also GNS3 are the simulators that I recommend.

With Python, NETWORKX can help

Node-Red, a visual programming library can be useful.

Agreed with the above statement.  Because IPv6 address space is 128-bits (2¹²⁸) in size, containing 340,282,366,920,938,463,463,374,607,431,768,211,456 IPv6 addresses.  This is significant increase from IPv4 which only contains 4,294,967,296 or 4.3 billion IPv4 addresses (IPv6 is 7.9×10²⁸ times as many as IPv4 addresses).

According to Gartner, only 6.4 billion connected "Things" of IoT will be used in 2016 whereby there is still plenty of IPv6 addresses to be used up for IoT purpose.

Yes, there are many IoT training offered by academic institutions as well as software & hardware products companies.  E.g., you can refer to the following YouTube link:

I never used it in this way personally but in my opinion it should be possible.

But I think it would be difficult to define what is the optimal placement of PVDGs.

Try this

Dear Tunde.

It depends on your research, normally you can check power signal strength, link quality, signal noise ratio, etc. There are many models are available to check the wireless LAN, you will find through the internet. 

Hi Asma,

Did you checked NetFlow or you can make your own Network Monitoring tool. Network Monitoring is  a Sampling technique it can be random or systematic so choose any sampling technique like I choose Network Monitoring using Poisson Sampling. And it also depends where you are performing your experiment on simulated environment(Mininet) or with real openflow enabled Switch.

Use google search engine and many reference works would pop up.

You're right that packet loss in wireless networks is more of an issue than it is with cabled networks. For this reason, IEEE 802.11 includes its own packet acknowledgment or retransmit protocol, at layer 2 (ARQ, or automatic repeat request), much as TCP does at layer 4. This is in addition to forward error correction schemes, such as Reed-Solomon and/or convolutional codes, that both wired and wireless networks would implement.

Which is why WiFi is not very good with IP multicast. The lower layer retries are incompatible with multicast, so they have to be disabled. Which leads to a relatively high level of missed packets.

Another possible technique, to distribute IP multicast streams to WiFi nets, is to convert the multicast packets into unicast packets, at the wired network edge. The problem is being considered by the IETF mboned working group.

Jet,

       This IEEE paper shows the tested energy efficiency of and 802.15.4 network during Tx, Rx, Ack, Ka, and Adv activity as well as during idle and off states: http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=6627960. Hopefully this paper offers you another option for calculating specific energy consumption on a TSCH network using the 802.15.4e (LR-WPAN) guidelines.

In addition to Imran Shafi and Jacek Bartman suggested parameters you have to also consider the values of mu and learning rate. More importantly training algorithm itself also has influence on error-different algorithm work differently with a particular problem.

The impact of DG on distribution system performance in general affected by many factors such as size, location, type, power factor ... of DG . for more details you can refer to my papers.

Dear

As I know the latency, packet loss/loss rate, throughput was occurred by the transfer data through the link and the root of the network.

So I prefer to do deference project using OPNET simulation with two scenario for each project once with IPv4 and the other with IPv6. You should use same configuration and load in each scenario of that project. Tray to do 3-6 projects with different traffic, nodes, link then you can analyze the result.

Thanks and Good luck .

QoS metrics could be deadline constraints and bandwidth it fully depends on system resources

Hello all,

actually, beside various BER estimation methods relying on assumptions about the statistical distribution of the signal characteristics (phase/amplitude) at the detector, VPItransmissionMaker Optical Systems also provides modules for estimating the BER based on error counting (noted as Monte-Carlo method).

As Mohamed pointed out above, you should make sure sufficient number of bits (or symbols if you use higher order modulation formats) is transmitted in order to count sufficient number of errors. As rough estimate, the accuracy of your estimated BER value scales with 1/sqrt(#-of-errors). So when you count 100 errors, you have an accuracy of approx. 1/sqrt(100) = 10%.

Best regards,

Andre

Hi Imed, Thanks once again for your reply. I am using framegrabber for processing image and then giving it to server laptop using RTSP. here for a heavy traffic i am using bitrate like 100kbps, 1mbps, 2mbps, 3mbps and 4 mbps. I have tested the system for all these rates but the problem is still same. 

Tried pinging back to back. Ping is available till I am asking for RTSP stream on client end. and then server going down after 30 seconds of start. I will try to check the repeated ping with -t option and will try to capture the ICMP Quench problem.

Your advice will be helpful for me to troubleshoot my problem in more detail.

Dear Lila,

It appears that you are overfitting the training data set. This is a very common problem when training neural networks, but more in general in machine learning. How to prevent overfitting is a large area of investigation. I would recommend you looking more in detail in machine learning and neural network literature. Common methods are called regularisation, early stop of training, reducing the number degree of freedom of the model, i.e. number of parameters, increasing the size of the data-set. 

TCP itself does not authenticate the receiver in any way. If the attacker replies on your TCP SYN-SENT faster than the intended socket for connection, the connection will be established with the attacker instead. That is basically the way some sort of Man-In-The-Middle attacks work. If you remember the Yahoo-highjacking by NSA last year - this worked alike (in that case according to the press even with full TSL Auth.: looking for request to Yahoo, changing some DNS servers to impose to be the receiver (the fake server was physically nearer to the sender and thus faster responding), issue a flasificated TLS Cert making sure the false Cert was verified faster than an original one (as above), and highjacking the connecting therefore.)

From wikipedia:

TCP-Handshake

1. SYN-SENT → <SEQ=100><CTL=SYN> → SYN-RECEIVED

2. SYN/ACK-RECEIVED ← <SEQ=300><ACK=101><CTL=SYN,ACK> ← SYN/ACK-SENT

3. ACK-SENT → <SEQ=101><ACK=301><CTL=ACK> → ESTABLISHED

Please provide sufficient information. It is not clear in what basis, you evaluated the results. What type of network is considered - is it wireless or wired?

Vincenzo, have a look on "journal of business venturing". At that place is plenty of research containing your interest. Also Per is on of the leading scholars in the field of entrepreneurship and start-ups.

This refers to the medium in which control traffic traverses. That is, which links are used to exchange control-related data between the controller and the forwarding devices it commands, or among controller instances.

Out-of-Band control uses separate Ethernet ports and links (i.e., a separate network) to connect forwarding devices to the controller and exchange control traffic. This separate infrastructure can be either physical or logical. In the first case, OpenFlow Switches have a physical "management" port which should be connected to controller or a traditional L2/L3 network connected to the controller(s) (this is traditional way used by todays' switch vendors). In the second case, tunneling or virtualization is used to create a separate logical network in the same physical infrastructure (e.g., in the original OpenFlow deploynment at Stanford, switches where configured with 3 VLANs, one for OpenFlow control traffic, one for experimenting data traffic with OpenFlow, and one for production traffic).

In-Band control uses the same links (i.e., same network) for both data traffic and control traffic. This case usually requires the switches to have a predefined set of rules for establishing the connection with the controller.

There is no much information about In-Band and Out-of-Band control in Software-Defined Networking. You can use the reference provided by Thomas Pfeiffenberger, as well as the OpenFlow and the OFConfig specifications to gain additional knowledge. Although, as far as I know, there is little information about it.

OPNET academic version does not come with LTE by default. Therefore you need to request for it to be added to OPNET modeler.

An approach that is focused on evolving the current protocols without a deeper change on them will be always limited to this environment.

A clean-slate architecture creates a new environment, with a new world of possibilities, not limited to the current ones.

Have a look at the downloads section, probably working ns-2 distributions are available there for multichannel support in wireless networks

There's also QoSMeT (Quality of Service Measurement Tool) from VTT Technical research centre of Finland. It allows you to measure two-way parameters (delay, jitter, throughput). In addition, VoIP MOS PSQA can calculated as well. It works in IP level, Win and Linux are supported. Great for pinpointing when locating bottlenecks in end-to-end connection. For academic partners, there is a free lisence. It is used by tens of universities in the world.

What might be of some interest of you is that especially in 2G environment operators are either blocking or hindering traffic to certain TCP ports (5060) in order to prevent VoIP, which might be cheaper than regular voice call (data tariffs dependent, of course).

Please go through tha paper entitled "Survey of Wireless Sensor Networks Simulation Tools for Demanding". Hope this paper will help you.

Hello,

I think this is a very interesting and challenging question!

I think that three will be the most challenging issues in 4G networks:

1. providing QoS for different service classes. Indeed, flows exchanged in 4G systems are characterized by different QoS requirements and by different priority levels;

2. unicast/multicast joint delivery. indeed, multicast services are expected to be massively exchanged in 4G networks, and their coexistence with traditional unicast flows poses additional scheduling issues.

3. scheduling in heterogeneous environments. Indeed, 4G networks offer different radio access technologies with heterogeneous base stations (macro, relay, femto, pico, etc. cells) and several transmissions types (cellular or direct device-to-device communications). All these features meaningfully increase the issues to solve in scheduling in 4G systems, making a given solution (e.g., cooperative or opportunistic) hard to be suitable in all the admissible scenarios.

I found these papers very interesting, i hope they could help you:

The most promising scheduling algorithm to provide guaranteed QoS to all types of traffic in multiservice 4G wireless networks

Joint Delivery of Unicast and E-MBMS Services in LTE Networks