Network Protocols - Science topic

David Lide Yeah but what are the protocols used to read those data. For an example FLPs are used to do auto negotiation. Like wise how are the other things are read or managed by the NIC?

A good way to start is to read recent overview articles and look for open problems that are stated in those papers. When it comes to machine learning, a general advice is to identify an existing protocol that you know is suboptimal, and then using machine learning to tweak the protocol.

Netsim simulator is the best, it's gathering all these features of all other simulators

Dear Colleagues:

Below is the up to date feasibility demonstration that you may like to have a look at.

Abe (2020-08-29 12:13 EDT)

Video transmission over wireless networks is considered the most interesting application in our daily life nowadays. This kind of application needs large bandwidth, efficient routing protocols, and content delivery methods to provide smooth video playback to the receivers. I hope you will benefit from this research paper

https://www.ripublication.com/ijeer17/ijeerv9n2_06.pdf

This research problem can not solve only by NS3 . You should use hybrid approach . For mobility and protocol implementation NS3 is good enough .

This study may help you out a bit .

10.1504/IJWMC.2019.10022912

Hi C K Gomathy Himadri Nath Saha ! Thank you for responding! Based on my research, a fuzzer may not trigger all possible behaviors of the program. One of the reasons is because it may require specific inputs to transverse certain code paths. Symbolic Execution or Concolic Execution is a method to analyze a program and generate specific inputs, with the purpose of having more code coverage (https://www.fuzzingbook.org/html/SymbolicFuzzer.html). I've asked this question because I am interested in using Symbolic Execution or Concolic Execution for my protocol-based fuzzer, thereby attempting to have more code coverage.

Assuming you mean "position" : there are currently no devices that claim to establish position of a vehicle at the nanometer, and no use case where this would be useful in the automotive or avionics domain.

Thanks Mr. Dia, and Mr.Tayal.

Zahid, I believe this article addresses the tradeoffs really well:

Distance vector protocols are simpler to understand and simpler to configure, but they are limited in the size of networks they will accommodate and the take longer to converge, compared with link state.

Most of these are 'simulated' attacks. It is unlikely a researcher can launch a 'real' attack against a network given the kind of security that enterprise class networks come with.

i am working on NS3

You may look at the following example (after downloading ns-3),

examples/wireless/wifi-simple-adhoc.cc

This example is mainly for ad-hoc network simulations.

This link might help you.. https://groups.google.com/forum/#!topic/ns-3-users/klAS-HqU92I and this tutorial:  https://www.nsnam.org/docs/release/3.14/tutorial/singlehtml/index.html#building-a-wireless-network-topology

Check out DMVPM Technology

The fact to define the complexity of the network from the perspective of the Network Administrator. For me I considered network complexity  as a tool or mechanism to ensure good security system when dealing with device configuration. A complex network architecture and implementation determine how securely you can relied on the  networks infrastructure against internal and external attacks. In most cases security mechanism in a network is not form with a simple rules when one wanted to make them both computationally and un-computationally secured. Hence, sometimes complexity of network can be seen from incapacity of one to understand or provide solution to the challenges faced by his networks and that can be  look from the knowledge based problem and experience  in the field. There so many ways to explore complexity of the network just like what "Rudra" has explained in above.

Hi,

There are - kinda. MPL (RFC7774) is a multicast routing protocol intended for the same type of deployment as RPL. We did some work on trickle-for-multicast-routing, see "Performance analysis of Trickle as a flooding mechanism", attached.

The IETF HOMENET wg published DNCP (RFC7787) which uses Trickle (RFC6206), and which itself is used by HNCP (RFC

To the best of my knowledge, Contiki/Cooja is the best for IoT considering most of IoT  protocols. in addition, it can be used as simulator or as emulator for test bed.

all the best

Hi , in Aquasim is very simple

My students are using COOJA, it is good

Dear Chandra,

If you deliver UDP packet of size N=1250 byets to the physical layer,there will be overhead which is added to the input data. If there is forward error correction, parity bits will be added to the input data according to used FEC encoder. The output rate of the FEC encoder will be Nx the code rate, where the code rate r fec = n/k is the ratio of the length of the output message to the the length of the input message.

During modulation you add the pilot carriers with specific date modulated upon them for channel estimation. These are overhead data with the rate say r pilot. In addition a cyclic prefix must be added to every OFDM symbol with the rate r cyclic prefix to overcome inter symbol interference..

These rates are standardized and are known. So the final size of the packet at the output of the physical layer= N(1 +  r fec + r pilot + r cyclic prefix.)

It is so that every network layer adds overhead to the the protocol data unit such that it can perform its function and for encapsulation.

You can find more information in the LINK:http://research.ijcaonline.org/volume64/number8/pxc3885415.pdf

and the link: www.cnrood.com/public/.../WiFi_Physical_Layer_and_Transm_Meas.pd...

Best wishes

It depends on the devices used to form a network where the DSR is running.  In a lab set up, suppose somebody builds a MANET using 14-15 inch laptops, the power consumption may range from 20-100 W for each laptop with an average 60 W. Cellphone chargers when charging a cellphone consume approximately 2-6 W with an average 3 W.

I suggest the following linkshttp://energyusecalculator.com/electricity_laptop.htm

If I got your question correctly, you are looking for topology discovery. If that's the case, I think you may have a look of these links. It might give u some ideas

1- http://blogs.ixiacom.com/ixia-blog/discovery-in-openflow-networks-at-onf-plugfest/

2- http://vlkan.com/blog/post/2013/08/06/sdn-discovery/

3- http://sdn-lab.com/2014/12/31/topology-discovery-with-ryu/

4- http://sdn-lab.com/2014/12/25/shortest-path-forwarding-with-openflow-on-ryu/

There is a paper called Efficient Topology Discovery in SDN as attached. 

Or maybe you can use multiple instances of iPerf to mimic different traffic types. Suppose that in Scenario 2 you create an iPerf instance in a station using UDP and certain bandwidth; another instance using TCP at another bandwidth and so on for as many Access Categories/Traffic Types as you want. This way you get the aggregate throughput from each iPerf instance and compare it to those achieved in Scenario 1 using the same tools (but instead of instances you run a single iPerf process in each station, of course).

QualNet is a commercial software. There is no free version. At most, you could ask for a 30-day trial license (http://web.scalable-networks.com/content/evaluation-programs).

Kari Sir,

I did check with the opencores but I could not. But I will check once again.

Thank you sir. if anyother input please mail me sir.

Take a look at using zeromq. You could effectively build up a multihop network on one device or across multiple devices using one of the models provided in zeromq. I have used zeromq to build simple multihop IoT networks within which I simulated/tested certain scenarios.

hwmp already includes AODV. This is to manage a wireless network with a wired backhaul.

Well MATLAB can be used for simulations. Other softwares include OMNET++, Castalia,

Castalia will work better. some people have used it.

Thank you for your response Arjav and Jan Marten. I think, I am working on the application layer. And I don't think that the data stream was encrypted.

I have used Networkminer as well, & it gives me detailed description of my captured data, in terms of images and other data format, but when it comes to video, they are shown as frames only! 

I have went back to Wireshark, used the export object option, looked into my http detail, but still no video is detected. What else should I do?!

This is highly determined by snooping configuration restrictions; where by on certain swithches, you may find that DHCP commands are not available. That is, if such are commands are entered, the switch gives an error message, hence configuration not applied. best way to go over this also is to look at the DHCP snooping configuration guidelines.

For educational purposes, please view this link.

@diamel teguig sir thnks for your view on it..

@muhammad naveed-sir How to determine the weight of a link in wireless mesh network and what are the criteria which kept in mind before allotting or assigning  weights to it..

Congratulations for your work and the nice approach to the Knapsack problem using agents. I see here an opportunity to improve the negotiation process between the Petitioner and Contributor Agents of my protocol (MOSAIC) using your strategies, and a framework for you (multilateral agreements) to extend your negotiation process involving a set of agents in the process and not only two. In addition to this, have you identified real use cases where this research could be deployed?

I am not sure if I understand, so, I think you need to first check the coverage. Secondly, it is interesting to do some layer 3 tests with sintetic traffic generators (like iperf), therefore, you can check throughput, packet loss, latency and other metrics.

Btw, the exchanged data is NOT encrypted with publick key. Public key crypto is used to create shared symetric keys by the handshake process in the beginning.

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

Well you may also want to add some routing. For example a default route might look like this

ipv6 route ::/0 2001:db8:01A5:C900::1

When reading a paper you should pay attention to the way the data was collected. Many of these "smooth curves" papers, as you mention, are theoretical ones. I mean, the results are obtained through the mathematical formulation/analysis of the problem. In such cases, the errors/variations play little role on the results, thus having more "harmonious" results. When analyzing real traffic logs, test bed implementation and even well conceived simulation/emulation experiments.... things are normally not very smooth. The amount of variables and different things that can impact the experiment rarely permits a very smooth output. However, you can minimize the discrepancies by:

- Run your experiment a fair amount of times (if in simulation with different seeds) so that it is statistically relevant. On the long run, if you have enough experiments, the spikes on one single experiment will have lower impact on the analysis of the whole experiment.

- Do a good statistical treatment of your data. Some statistical tools like percentile, weighted mean, mode, provide the correlation coefficient and confidence intervals... all these and other statistical methods/tools can, not only improve the quality of the results but help you (and your readers) to understand your results.

- Most of all, do not forget that is your job to analyze your graphs and drawn conclusions from it, not from your reader. YOU have access to all the codes, raw data and you have more time to analyze and understand your results. The reader should not do this job for you. If you are not satisfied with your results.... imagine the reader. Always put yourself on the shoes of the reader and see if you would be satisfied by the quality of the graph, what it is showing and the explanations provided for "strange" behaviors.

As far as I understood, you are not interesting in one specific technology are you? If it is just a theoretical question, and you are proposing a new mac layer. The answer would be yes, and yes you could implement that in ns-2 (even though I would urge you to use some other, more recent simulator). If you do not want to go into the internals of the PHY, what can be quite challenging if you do not know the PHY, or even if kind of nonsense, if you do not care for one specific technology. The eaisest would just add a flag on your packet with the specific frequency the source is using, and in the receiver ignore all the messages that does not match the frequency of the receiver.... simple like that.

OK that is cheating, surely it is, but if you are not interested in the phy, what you are trying to show is (protocol) is uper layer related... lets say topology control, group communication, filtering of data dissemination.... that is more than enough and quite easy to implement in almost ANY network simulator.

You can have a look at

Hi,

yes they are similar concepts. However the proper terms are: Master/Slaves and Farmer/Workers. Both refer to a model of parallelism in which the Master (or the Farmer) orchestrate the work to be done by Slaves (or Workers).

Usually the Master (Farmer) performs the initial calculations and then starts a loop in which the computational workload is assigned to the available Slaves (Workers).

Kind Regards,

--Osvaldo

you're right Imegh but time is an important factor in message communication (specially safety messages exchanging in VANET) and full encryption is a time consuming process that where possible, we should avoid it.

So, you can do one thing,first decide on which protocol you want to work and then download its files and refer on which configuration it will work.