Final Report Summary - INTENT (Information-Centric Network Management and Traffic Engineering)
Information-Centric Network Management and Traffic Engineering (INTENT)
The current Internet architecture focuses on communicating entities, leaving aside the exchanged information. However, current trends show that what is exchanged is becoming more important than who is exchanging it. As a result, the Internet is effectively moving from interconnecting machines to interconnecting information. Moreover, the location-centric model used by the underlying Internet routing paradigm, i.e. transferring datagrams from one routable endpoint to another, limits the ability to fully utilize resources that are available along the route from the provider to the consumer(s), such as storage or computing. Information-Centric Networking (ICN) has been proposed as a paradigm shift from the host-to-host Internet to a host-to content one, or in other words from an end-to-end communication system to a native distribution network. This trend has attracted the attention of the research community, which has argued that content, instead of end-points, must be at the centre stage of attention. These efforts promise, among other things, more flexibility in adapting to new services, efficiency improvements in lower layers, and native multicast support. However, crucial questions still remain, such as the management of ICN networks which has not yet been addressed focusing (1) on Traffic Engineering (TE) and (2) cache and route management. Thus, the main goals of the project “Information-Centric Network Management and Traffic Engineering” (INTENT) was to develop key network management functions for ICN networks related to route and cache management and to specify an overall ICN TE framework for content delivery.
The INTENT project developed an intelligent content management substrate which has distributed functionality across an ICN network with caching capabilities. This substrate enabled the coordination of key ICN management functionalities for achieving traffic engineering objectives pertaining to: (a) the placement of content at the various caching locations, and (b) the selection of the appropriate caching point(s) from which content requests are subsequently served. Since the above decisions can affect the performance of network resources, INTENT also investigated methods through which these can be jointly optimised. In addition to the static allocation of caches to user requests, INTENT also considered a more dynamic approach by which a user can be redirected to another cache, holding a replica of the requested content when the user experience falls below an acceptable level or the network experiences various link or node failures. The content placement and cache selection algorithms developed by INTENT already addressed shortcomings of current approaches in terms of processing and communication overhead, which allows their execution in short timescales, enabling them to react to changing characteristics of content popularity and demand as well as topological changes. In addition, their distributed nature manages to avoid inherent problems of centralised solutions such as the single point of failure. The proposed content delivery models are not necessarily relying on CDN providers. INTENT focused on empowering ISPs with in-network caching capabilities, which allows them to have more control over their network resources by implementing their own content placement and server selection strategies. Unlike the heavyweight caching infrastructure maintained by CDN providers, the INTENT solution allows ISPs to operate a simpler and more lightweight content delivery service, which will be of lower economic cost given that the distributed caching capability could be realised by commodity hardware components (e.g. small caching units at the edges of the network or attached at the core routers). The ICN paradigm moves away INTENT's TE logic from classical data path oriented TE techniques. It inherently provides a wider potential for maximising the utilisation of planned resources in comparison to the traditional shortest path routing schemes. All these constitute clear innovations of the INTENT project.
The main objectives of INTENT that were fully fulfilled are the following:
Objective 1: Specify an overall, technology-agnostic, ICN traffic engineering framework, encompassing the necessary management and control functions for caching and routing, so that to ensure that content is delivered as required, while network performance is within acceptable levels.
Objective 2: Design efficient caching algorithms deciding on the network location of caches, what objects to cache, influence the cache replacement policies, investigating both opportunistic in-network caching as well as CDN-like replication.
Objective 3: Design mechanisms and algorithms that manage the routing processes of ICNs by influencing the forwarding of interest/subscription packets optimising the network performance in terms of network load, congestion and delay.
Objective 4: Realise the specified ICN TE framework over both evolutionary IP-based ICN environments or over clean-slate ICN infrastructures such as CCN/NDN, including extensions to the architectures themselves.
Objective 5: Develop instances of experimental testbeds and various event-driven simulators. The testbeb setup and the used/extended simulators were used for the evaluation of the functionality and performance of the mechanisms and algorithms developed in the rest of the objectives, assessing the overall benefits of the INTENT project.
During the duration of the fellowship the work matching its objectives resulted in the publication of 8 peer reviewed conference/workshop papers and 5 peer reviewed top tier journal papers (IEEE Communication Magazine, IEEE JSAC, IEEE TWC, Elsevier Computer Networks and Elsevier Computer Communications). One more journal paper has received positive reviews and is in the process of major revisions (Elsevier Computer Networks). Finally, the researcher has under submission two more conference papers and one more journal paper (IEEE TNSM). Overall a total of 10 workshop/conference papers and 7 journal papers is the outcome of the work, whereas more submission and extensions are in preparation and due for submission immediately after the end of the project. Overall, all the milestones set for the fellowship have been successfully reached. The success of the fellowship and the produced research outcome is also reinforced by the following awards:
• The work: Vasilis Sourlas, Leandros Tassiulas, Ioannis Psaras and George Pavlou, “Information Resilience through User-Assisted Caching in Disruptive Content-Centric Networks” in 14th IFIP NETWORKING, Toulouse, France, May 2015 received the BEST PAPER AWARD of the conference.
• The work: Onur Ascigil, Vasilis Sourlas, Ioannis Psaras and George Pavlou, “Opportunistic Off-Path Content Discovery in Information-Centric Networks”, in IEEE LANMAN, Rome, Italy, June 2016 received the BEST PAPER AWARD of the conference.
• The work: Ioannis Psaras, Sergi Rene, Konstantinos V. Katsaros, Nikolaos Bezirgiannidis, Vasilis Sourlas, Sotiris Diamantopoulos, Ioannis Komnios, Vassilis Tsaoussidis and George Pavlou, "Keyword-based Mobile Application Sharing," in ACM MobiArch, 2016, New York, USA, Oct. 2016 received the BEST PAPER AWARD of the conference.
Besides the publications, the researcher participated and presented the results of INTENT in 4 IETF ICNRG meetings (IETF 93, IETF 94, Paris Interim meeting Jan. 2015 and IETF 96). He also gave 12 invited talks is various universities and research labs in Europe, US and Japan.
More details about the outcome of the INTENT can be found in the project’s web page:
https://sites.google.com/site/mciefintentfp7/
as well as the personal web pages of the:
1) Scientist in charge: Prof. George Pavlou, https://www.ee.ucl.ac.uk/~gpavlou/
2) Fellow researcher: Dr. Vasilis Sourlas, https://sites.google.com/site/vsourlas/Home
The current Internet architecture focuses on communicating entities, leaving aside the exchanged information. However, current trends show that what is exchanged is becoming more important than who is exchanging it. As a result, the Internet is effectively moving from interconnecting machines to interconnecting information. Moreover, the location-centric model used by the underlying Internet routing paradigm, i.e. transferring datagrams from one routable endpoint to another, limits the ability to fully utilize resources that are available along the route from the provider to the consumer(s), such as storage or computing. Information-Centric Networking (ICN) has been proposed as a paradigm shift from the host-to-host Internet to a host-to content one, or in other words from an end-to-end communication system to a native distribution network. This trend has attracted the attention of the research community, which has argued that content, instead of end-points, must be at the centre stage of attention. These efforts promise, among other things, more flexibility in adapting to new services, efficiency improvements in lower layers, and native multicast support. However, crucial questions still remain, such as the management of ICN networks which has not yet been addressed focusing (1) on Traffic Engineering (TE) and (2) cache and route management. Thus, the main goals of the project “Information-Centric Network Management and Traffic Engineering” (INTENT) was to develop key network management functions for ICN networks related to route and cache management and to specify an overall ICN TE framework for content delivery.
The INTENT project developed an intelligent content management substrate which has distributed functionality across an ICN network with caching capabilities. This substrate enabled the coordination of key ICN management functionalities for achieving traffic engineering objectives pertaining to: (a) the placement of content at the various caching locations, and (b) the selection of the appropriate caching point(s) from which content requests are subsequently served. Since the above decisions can affect the performance of network resources, INTENT also investigated methods through which these can be jointly optimised. In addition to the static allocation of caches to user requests, INTENT also considered a more dynamic approach by which a user can be redirected to another cache, holding a replica of the requested content when the user experience falls below an acceptable level or the network experiences various link or node failures. The content placement and cache selection algorithms developed by INTENT already addressed shortcomings of current approaches in terms of processing and communication overhead, which allows their execution in short timescales, enabling them to react to changing characteristics of content popularity and demand as well as topological changes. In addition, their distributed nature manages to avoid inherent problems of centralised solutions such as the single point of failure. The proposed content delivery models are not necessarily relying on CDN providers. INTENT focused on empowering ISPs with in-network caching capabilities, which allows them to have more control over their network resources by implementing their own content placement and server selection strategies. Unlike the heavyweight caching infrastructure maintained by CDN providers, the INTENT solution allows ISPs to operate a simpler and more lightweight content delivery service, which will be of lower economic cost given that the distributed caching capability could be realised by commodity hardware components (e.g. small caching units at the edges of the network or attached at the core routers). The ICN paradigm moves away INTENT's TE logic from classical data path oriented TE techniques. It inherently provides a wider potential for maximising the utilisation of planned resources in comparison to the traditional shortest path routing schemes. All these constitute clear innovations of the INTENT project.
The main objectives of INTENT that were fully fulfilled are the following:
Objective 1: Specify an overall, technology-agnostic, ICN traffic engineering framework, encompassing the necessary management and control functions for caching and routing, so that to ensure that content is delivered as required, while network performance is within acceptable levels.
Objective 2: Design efficient caching algorithms deciding on the network location of caches, what objects to cache, influence the cache replacement policies, investigating both opportunistic in-network caching as well as CDN-like replication.
Objective 3: Design mechanisms and algorithms that manage the routing processes of ICNs by influencing the forwarding of interest/subscription packets optimising the network performance in terms of network load, congestion and delay.
Objective 4: Realise the specified ICN TE framework over both evolutionary IP-based ICN environments or over clean-slate ICN infrastructures such as CCN/NDN, including extensions to the architectures themselves.
Objective 5: Develop instances of experimental testbeds and various event-driven simulators. The testbeb setup and the used/extended simulators were used for the evaluation of the functionality and performance of the mechanisms and algorithms developed in the rest of the objectives, assessing the overall benefits of the INTENT project.
During the duration of the fellowship the work matching its objectives resulted in the publication of 8 peer reviewed conference/workshop papers and 5 peer reviewed top tier journal papers (IEEE Communication Magazine, IEEE JSAC, IEEE TWC, Elsevier Computer Networks and Elsevier Computer Communications). One more journal paper has received positive reviews and is in the process of major revisions (Elsevier Computer Networks). Finally, the researcher has under submission two more conference papers and one more journal paper (IEEE TNSM). Overall a total of 10 workshop/conference papers and 7 journal papers is the outcome of the work, whereas more submission and extensions are in preparation and due for submission immediately after the end of the project. Overall, all the milestones set for the fellowship have been successfully reached. The success of the fellowship and the produced research outcome is also reinforced by the following awards:
• The work: Vasilis Sourlas, Leandros Tassiulas, Ioannis Psaras and George Pavlou, “Information Resilience through User-Assisted Caching in Disruptive Content-Centric Networks” in 14th IFIP NETWORKING, Toulouse, France, May 2015 received the BEST PAPER AWARD of the conference.
• The work: Onur Ascigil, Vasilis Sourlas, Ioannis Psaras and George Pavlou, “Opportunistic Off-Path Content Discovery in Information-Centric Networks”, in IEEE LANMAN, Rome, Italy, June 2016 received the BEST PAPER AWARD of the conference.
• The work: Ioannis Psaras, Sergi Rene, Konstantinos V. Katsaros, Nikolaos Bezirgiannidis, Vasilis Sourlas, Sotiris Diamantopoulos, Ioannis Komnios, Vassilis Tsaoussidis and George Pavlou, "Keyword-based Mobile Application Sharing," in ACM MobiArch, 2016, New York, USA, Oct. 2016 received the BEST PAPER AWARD of the conference.
Besides the publications, the researcher participated and presented the results of INTENT in 4 IETF ICNRG meetings (IETF 93, IETF 94, Paris Interim meeting Jan. 2015 and IETF 96). He also gave 12 invited talks is various universities and research labs in Europe, US and Japan.
More details about the outcome of the INTENT can be found in the project’s web page:
https://sites.google.com/site/mciefintentfp7/
as well as the personal web pages of the:
1) Scientist in charge: Prof. George Pavlou, https://www.ee.ucl.ac.uk/~gpavlou/
2) Fellow researcher: Dr. Vasilis Sourlas, https://sites.google.com/site/vsourlas/Home