GSOC2025Projects - Revision history

Tomh: /* Linux-like Loss Detection Techniques for ns-3 TCP */

2025-10-24T19:49:28Z

Linux-like Loss Detection Techniques for ns-3 TCP

← Older revision		Revision as of 19:49, 24 October 2025
Line 255:		Line 255:
	* ''Difficulty:'' Medium to Hard.		* ''Difficulty:'' Medium to Hard.
	* ''Recommended Reading:''		* ''Recommended Reading:''
	** Forward Acknowledgement (FACK) [[https://~~citeseerx~~.~~ist~~.~~psu.edu~~/~~viewdoc~~/~~download?doi=~~10.~~1.1.41~~.~~2559&rep=rep1&type=pdf~~ Paper]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]		** Forward Acknowledgement (FACK) [[https://dl.acm.org/doi/abs/10.1145/248157.248181 Paper]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]
	** Duplicate Selective Acknowledgment (DSACK) [[https://tools.ietf.org/html/rfc2883 RFC 2883]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]		** Duplicate Selective Acknowledgment (DSACK) [[https://tools.ietf.org/html/rfc2883 RFC 2883]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]
	** RACK-TLP [[https://datatracker.ietf.org/doc/html/rfc8985 RFC 8985]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]		** RACK-TLP [[https://datatracker.ietf.org/doc/html/rfc8985 RFC 8985]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]

Tommaso: /* IPv6 global routing */

2025-04-19T15:56:46Z

IPv6 global routing

← Older revision		Revision as of 15:56, 19 April 2025
Line 284:		Line 284:
	=== IPv6 global routing ===		=== IPv6 global routing ===

	Mentors: [mailto:tommaso.pecorella@unifi.it Tommaso Pecorella], [~~TBD~~].		Mentors: [mailto:tommaso.pecorella@unifi.it Tommaso Pecorella], [mailto:manoj24.rana@gmail.com Manoj K. Rana].

	Creating a complex topology can be a problem, and sometimes the user do not want to be (also) concerned about setting up dynamic routing protocols (e.g., RIP, RIPng).		Creating a complex topology can be a problem, and sometimes the user do not want to be (also) concerned about setting up dynamic routing protocols (e.g., RIP, RIPng).

Tommaso: /* 6LoWPAN neighbor discovery protocol */

2025-04-19T15:55:40Z

6LoWPAN neighbor discovery protocol

← Older revision		Revision as of 15:55, 19 April 2025
Line 197:		Line 197:
	=== 6LoWPAN neighbor discovery protocol ===		=== 6LoWPAN neighbor discovery protocol ===

	Mentors: [mailto:tommaso.pecorella@unifi.it Tommaso Pecorella], [~~TBD~~].		Mentors: [mailto:tommaso.pecorella@unifi.it Tommaso Pecorella], [mailto:adnan.rashid@poliba.it Adnan Rashid].

	The 6LoWPAN-ND (RFCs 4944, 6775, and 8505) is a replacement for IPv6 DAD and NDP for 6LoWPAN networks, and it is important to ensure address uniquness across a network that can potentially use different MAC/PHY layers.		The 6LoWPAN-ND (RFCs 4944, 6775, and 8505) is a replacement for IPv6 DAD and NDP for 6LoWPAN networks, and it is important to ensure address uniquness across a network that can potentially use different MAC/PHY layers.

Tommaso: /* 6LoWPAN mesh-under routing enhancements */

2025-03-17T17:22:25Z

6LoWPAN mesh-under routing enhancements

← Older revision		Revision as of 17:22, 17 March 2025
Line 193:		Line 193:

	Possible tasks to fulfill the patch requirement:		Possible tasks to fulfill the patch requirement:
	* ~~TBD~~		* The current code is hardwired, i.e., the mesh-under routing scheme is embedded into the SixLowPanNetDevice. Propose a patch to decouple it, using a SixLowPanMeshUnderRouting class to determine the "next hop". The behavior of the actual protocol should be unchanged.

	=== 6LoWPAN neighbor discovery protocol ===		=== 6LoWPAN neighbor discovery protocol ===

Biljana: /* 5G NR module integration with ns-3-ai */

2025-03-07T14:45:32Z

5G NR module integration with ns-3-ai

← Older revision		Revision as of 14:45, 7 March 2025
Line 353:		Line 353:
	Mentors: [mailto:kkoutlia@cttc.es Katerina Koutlia], [mailto:gcarvalho@cttc.es Gabriel Ferreira], [mailto:aashtari@cttc.es Amir Ashtari Gargari] and [mailto:bbojovic@cttc.es Biljana Bojovic]		Mentors: [mailto:kkoutlia@cttc.es Katerina Koutlia], [mailto:gcarvalho@cttc.es Gabriel Ferreira], [mailto:aashtari@cttc.es Amir Ashtari Gargari] and [mailto:bbojovic@cttc.es Biljana Bojovic]

	The objective of this project is to integrate the ns-3 5G NR module with [https://apps.nsnam.org/app/ns3-ai/ ns-3-ai]. In GSoC 2024 we had a project in which 5G NR was integrated with [https://www.nsnam.org/wiki/GSOC2024RLUsability5G ns-3 gym]. While ns-3 gym is a popular ns-3 module for AI, it is limited to the application of reinforcement ~~learning machine~~ learning techniques in networking research. On the other hand, ns-3-ai module provides a more general solution that enables the data interaction between ns-3 and other Python-based AI frameworks, like [https://www.tensorflow.org/api_docs/cc Tensorflow C++ APIs] and [https://pytorch.org/cppdocs/ PyTorch C++ APIs], which opens the door to ~~the~~ use different machine learning-based techniques in 5G NR models. The correct functioning of the integration should be tested, and documented, and a fully working example using ns-3-ai should be provided. The contributor can propose a use-case scenario for matching learning. One option is to use it for MAC scheduling, but it could be used for other research problems, and the contributor is encouraged to propose the use case of his/her interest.		The objective of this project is to integrate the ns-3 5G NR module with [https://apps.nsnam.org/app/ns3-ai/ ns-3-ai]. In GSoC 2024 we had a project in which 5G NR was integrated with [https://www.nsnam.org/wiki/GSOC2024RLUsability5G ns-3 gym]. While ns-3 gym is a popular ns-3 module for AI, it is limited to the application of reinforcement learning techniques in networking research. On the other hand, ns-3-ai module provides a more general solution that enables the data interaction between ns-3 and other Python-based AI frameworks, like [https://www.tensorflow.org/api_docs/cc Tensorflow C++ APIs] and [https://pytorch.org/cppdocs/ PyTorch C++ APIs], which opens the door to use different machine learning-based techniques in 5G NR models. The correct functioning of the integration should be tested, and documented, and a fully working example using ns-3-ai should be provided. The contributor can propose a use-case scenario for matching learning. One option is to use it for MAC scheduling, but it could be used for other 5G related research problems, and the contributor is encouraged to propose the use case of his/her interest.

	For starters, we would suggest adding the CTTC 5G-LENA (nr module) to ns-3, in the typical way (as a module in the contrib/ directory), and building and running the examples. Documentation is available from here: https://5g-lena.cttc.es/. There is an overview tutorial video available here: https://acmse.net/2021/tutorials-offered/#tut-work03. That is the background information.		For starters, we would suggest adding the CTTC 5G-LENA (nr module) to ns-3, in the typical way (as a module in the contrib/ directory), and building and running the examples. Documentation is available from here: https://5g-lena.cttc.es/. There is an overview tutorial video available here: https://acmse.net/2021/tutorials-offered/#tut-work03. That is the background information.

Tomh: /* Switched Ethernet */

2025-03-02T14:21:14Z

Switched Ethernet

← Older revision		Revision as of 14:21, 2 March 2025
Line 398:		Line 398:
	** [https://www2.nsnam.org/bugzilla/show_bug.cgi?id=2602 Basic implementation (very old)]		** [https://www2.nsnam.org/bugzilla/show_bug.cgi?id=2602 Basic implementation (very old)]
	** [https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/431 Full-duplex CSMA, somehow related]		** [https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/431 Full-duplex CSMA, somehow related]
			** [https://tcipg.org/sites/default/files/papers/2010_Jin_Nicol_Caesar.pdf Paper on gigabit Ethernet switch models for simulation]

	Possible tasks to fulfill the patch requirement:		Possible tasks to fulfill the patch requirement:
	* TBD, contact the mentors if interested.		* TBD, contact the mentors if interested.

Tommaso: /* Internet models enhancements */

2025-02-28T19:39:45Z

Internet models enhancements

← Older revision		Revision as of 19:39, 28 February 2025
Line 79:		Line 79:
	* [[#AODVv2 Protocol enhancements]] (medium size project, 175h)		* [[#AODVv2 Protocol enhancements]] (medium size project, 175h)
	* [[#IPv6 global routing]] (large size project, 350h)		* [[#IPv6 global routing]] (large size project, 350h)
	* [[#IPv6 support in Ad hoc Routing Protocols]] (large size project, 350h)
	* [[#Linux-like CAKE queue discipline for ns-3]] (large size project, 350h)		* [[#Linux-like CAKE queue discipline for ns-3]] (large size project, 350h)
	* [[#Switched Ethernet]] (large size project, 350h)		* [[#Switched Ethernet]] (large size project, 350h)

Tommaso: /* Internet models enhancements */

2025-02-16T16:48:07Z

Internet models enhancements

← Older revision		Revision as of 16:48, 16 February 2025
Line 81:		Line 81:
	* [[#IPv6 support in Ad hoc Routing Protocols]] (large size project, 350h)		* [[#IPv6 support in Ad hoc Routing Protocols]] (large size project, 350h)
	* [[#Linux-like CAKE queue discipline for ns-3]] (large size project, 350h)		* [[#Linux-like CAKE queue discipline for ns-3]] (large size project, 350h)
			* [[#Switched Ethernet]] (large size project, 350h)

	=== IoT models enhancements ===		=== IoT models enhancements ===

Tommaso: /* Large projects (350 hours) */

2025-02-16T16:47:12Z

Large projects (350 hours)

← Older revision		Revision as of 16:47, 16 February 2025
Line 377:		Line 377:
	** [https://lwn.net/Articles/758353/ Let them run CAKE]		** [https://lwn.net/Articles/758353/ Let them run CAKE]
	** [https://www.nsnam.org/docs/models/html/queue-discs.html Queue disciplines in ns-3]		** [https://www.nsnam.org/docs/models/html/queue-discs.html Queue disciplines in ns-3]

			=== Switched Ethernet ===

			Mentors: [mailto:tommaso.pecorella@unifi.it Tommaso Pecorella], TBD.

			The current ns-3 models for wired connections are fine for simple networks, but the lack of a switched Ethernet model is a limitation in some cases.

			The goal of the idea is to create, test, and document a Switched Ethernet model, able to simulate (at least) 1, 10, and 40 GbE links and model for a switch.

			The model of the NetDevice and Channel shall take into account the link delays and errors, similarly to what is done by the point-to-point model. Futhermore, it should be able to set the link speed and if it is full-duplex or half-duplex. Additional support for flow control is a bonus, but not strictly required. Link speed auto-negotiation is not considered to be interesting.

			The model for the switch should be modular (i.e., allowing the development of different switch types), and include auto-learning of I/O ports based on the MAC address, i.e., have a MAC/port table, and a basic store-and-forward operation. Features like advanced I/O buffer handling and ARP/NDP spoofing detection are not a priority and shall be left for future implementations.

			The model should consider the future implementaion of algorithms like VLANs (IEEE 802.1Q, 802.1ad), and the Spanning Tree Protocol (IEEE 802.1D, 802.1w, and 802.1s). Their implementaion is not required, but the model design should allow their development.

			* ''Required Experience:'' Fundamentals of Ethernet sitched networking, C++ programming.
			* ''Interests:'' Ethernet networks and switched data networks.
			* ''Difficulty:'' Medium.
			* ''Recommended reading:''
			** [https://www2.nsnam.org/bugzilla/show_bug.cgi?id=2602 Basic implementation (very old)]
			** [https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/431 Full-duplex CSMA, somehow related]

			Possible tasks to fulfill the patch requirement:
			* TBD, contact the mentors if interested.

Tommaso: /* Medium sized projects (175 hours) */

2025-02-16T16:06:49Z

Medium sized projects (175 hours)

@@ Line 258: / Line 258: @@
 ** Duplicate Selective Acknowledgment (DSACK) [[https://tools.ietf.org/html/rfc2883 RFC 2883]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]
 ** RACK-TLP [[https://datatracker.ietf.org/doc/html/rfc8985 RFC 8985]] [[https://gitlab.com/nsnam/ns-3-dev/-/merge_requests/34/commits Implementation]]
 == Large projects (350 hours) ==