INSP: The business and technical case for In-Network Service Providers
Lead Research Organisation:
University College London
Department Name: Electronic and Electrical Engineering
Abstract
This project will design and implement all the required functionality to de-ossify the Internet architecture, which at the moment is limited to end-host-centric communication realised through end-host IP addresses. The IP protocol, which has become the thin-waist of the protocol stack, on the one hand guarantees a common language for communication between any network and any device, but on the other, restricts both communication and computation at the end-points of any connection. This further limits the Internet architecture from integrating new functionalities at the core of the network, unavoidably pushing innovation at the application layer.
We argue that network functionality (in terms of content and service management) does not necessarily need to be pushed at the application-layer of the connection's end-points, but can instead be managed at the network-layer of mid-path network entities.
To achieve our goals we will make use of the basic principles of two new networking paradigms, namely Information-Centric Networks (ICNs) and Software-Defined Networks (SDNs). ICNs introduce the notion of named-content objects and enable management of named-content at the network-layer (or data-plane) of any network device. SDNs on the other hand, focus on network management operations and push functionality at the control plane. Although both paradigms hold huge potential to change internetworking as we know it today, they both face implementation challenges.
In this project, we will combine the functionality of ICNs and SDNs and take advantage of their novel concepts to design a new common Internet playground, where innovation will not be forcefully pushed at the application layer. We will follow pragmatic implementation routes to guarantee that the challenges faced in the areas of ICN and SDN are smoothly overcome.
The high-level objectives of the project are the following:
1. implement and integrate the required technology to allow Internet Service Providers (ISPs) implement their own protocols of choice within their administrative domains, allowing for de-ossification of the Internet architecture.
2. allow in-network entities to execute logic and run instances of applications mid-path from source to destination, which are otherwise run within data-centres, more often than not, far away from the end-users.
The above targets will be achieved by introducing a new Internet market player, which we call In-Network Service Provider (or INSP) and who will take on the task (and cost) of implementing the proposed functionality.
We argue that network functionality (in terms of content and service management) does not necessarily need to be pushed at the application-layer of the connection's end-points, but can instead be managed at the network-layer of mid-path network entities.
To achieve our goals we will make use of the basic principles of two new networking paradigms, namely Information-Centric Networks (ICNs) and Software-Defined Networks (SDNs). ICNs introduce the notion of named-content objects and enable management of named-content at the network-layer (or data-plane) of any network device. SDNs on the other hand, focus on network management operations and push functionality at the control plane. Although both paradigms hold huge potential to change internetworking as we know it today, they both face implementation challenges.
In this project, we will combine the functionality of ICNs and SDNs and take advantage of their novel concepts to design a new common Internet playground, where innovation will not be forcefully pushed at the application layer. We will follow pragmatic implementation routes to guarantee that the challenges faced in the areas of ICN and SDN are smoothly overcome.
The high-level objectives of the project are the following:
1. implement and integrate the required technology to allow Internet Service Providers (ISPs) implement their own protocols of choice within their administrative domains, allowing for de-ossification of the Internet architecture.
2. allow in-network entities to execute logic and run instances of applications mid-path from source to destination, which are otherwise run within data-centres, more often than not, far away from the end-users.
The above targets will be achieved by introducing a new Internet market player, which we call In-Network Service Provider (or INSP) and who will take on the task (and cost) of implementing the proposed functionality.
Planned Impact
This is a high-risk/high-impact project. The timescale of the projected impact is targeting the period of the next 10-30 years. In case of successful adoption of the project results, which we are confident will be significant, network operation and application execution will be revolutionised.
Internet Service Providers (ISPs) will have the chance to administer their domains according to demand-supply ratios, rather than being passive "bit pushers", struggling to keep their networks running, as is the case today. ISPs operating in rural areas will have the chance to operate under simpler intra-domain routing protocols for instance, or applying different traffic engineering techniques. In contrast, a mobile operator in a very densely populated area, will need more scalable mobile communication protocols. For example, TCP has been repeatedly shown to perform poorly under mobility conditions and although dozens of alternatives exist, they cannot inter-operate with conventional TCP, due to the highly rigid Internet core functionality.
Application logic will not need to take place in massive-scale data-centres only. Instead, instances or fractions of applications for peer-assisted services will also execute mid-path from source to destination. This alters application development to run in a distributed manner and opens radical new avenues for unforeseen applications.
Our handheld devices (smartphones) today are more powerful than the first computers were 40 years ago (which occupied whole buildings, similarly to today's data-centres). The Internet architecture has to be ready to integrate the data-centres of the future and exploit resources in the best possible manner. This project sets out to be a stepping stone towards this direction. The project anticipates to open a new research area, in which the academic community will have to engage and take further.
Given the forward-looking and ambitious vision of this project, whose result is expected to show within the 10 to 30 year timeframe, we split the expected impact in short-, medium- and long-term.
Short-term Impact (0-5 years):
This time period is covering the lifetime of the project and we expect the main benefit here to come in terms of research outcome. That said, the academic community will benefit from the new research area we are seeking to establish.
Medium-term Impact (5-10 years):
This time period will follow the completion of the project and is the period during which research results will be adjusted to the needs of real implementations. Although during the project the design will be based on realistic settings, hardware and software technology evolution will influence implementation settings. We expect that both the academic community and the R&D industry will work on adapting the solutions of the project to real-world environments and guarantee full backwards compatibility.
Furthermore, new research areas will be formed pushing innovation inside the networking core, rather than at the application layer.
Long-term Impact (10-30 years):
We expect the following entities to benefit from the successful adoption of the results of this project in the long term:
- ISPs: will have the opportunity to actively manage traffic in their networks according to supply and demand and will be able to operate according to their own protocols and traffic engineering rules.
- Application developers: will have the opportunity to run instances of their applications in a distributed manner, closer to the audience they are targeting and not in a centralised fashion within data-centres. This will also reduce their costs, as they won't need to pay big cloud providers with global footprint, but rather local INSPs, where their audience is based. It will also help smaller companies and individuals to get into the Internet market area.
- Users: will see improved performance in terms of content delivery and application or service response times.
Internet Service Providers (ISPs) will have the chance to administer their domains according to demand-supply ratios, rather than being passive "bit pushers", struggling to keep their networks running, as is the case today. ISPs operating in rural areas will have the chance to operate under simpler intra-domain routing protocols for instance, or applying different traffic engineering techniques. In contrast, a mobile operator in a very densely populated area, will need more scalable mobile communication protocols. For example, TCP has been repeatedly shown to perform poorly under mobility conditions and although dozens of alternatives exist, they cannot inter-operate with conventional TCP, due to the highly rigid Internet core functionality.
Application logic will not need to take place in massive-scale data-centres only. Instead, instances or fractions of applications for peer-assisted services will also execute mid-path from source to destination. This alters application development to run in a distributed manner and opens radical new avenues for unforeseen applications.
Our handheld devices (smartphones) today are more powerful than the first computers were 40 years ago (which occupied whole buildings, similarly to today's data-centres). The Internet architecture has to be ready to integrate the data-centres of the future and exploit resources in the best possible manner. This project sets out to be a stepping stone towards this direction. The project anticipates to open a new research area, in which the academic community will have to engage and take further.
Given the forward-looking and ambitious vision of this project, whose result is expected to show within the 10 to 30 year timeframe, we split the expected impact in short-, medium- and long-term.
Short-term Impact (0-5 years):
This time period is covering the lifetime of the project and we expect the main benefit here to come in terms of research outcome. That said, the academic community will benefit from the new research area we are seeking to establish.
Medium-term Impact (5-10 years):
This time period will follow the completion of the project and is the period during which research results will be adjusted to the needs of real implementations. Although during the project the design will be based on realistic settings, hardware and software technology evolution will influence implementation settings. We expect that both the academic community and the R&D industry will work on adapting the solutions of the project to real-world environments and guarantee full backwards compatibility.
Furthermore, new research areas will be formed pushing innovation inside the networking core, rather than at the application layer.
Long-term Impact (10-30 years):
We expect the following entities to benefit from the successful adoption of the results of this project in the long term:
- ISPs: will have the opportunity to actively manage traffic in their networks according to supply and demand and will be able to operate according to their own protocols and traffic engineering rules.
- Application developers: will have the opportunity to run instances of their applications in a distributed manner, closer to the audience they are targeting and not in a centralised fashion within data-centres. This will also reduce their costs, as they won't need to pay big cloud providers with global footprint, but rather local INSPs, where their audience is based. It will also help smaller companies and individuals to get into the Internet market area.
- Users: will see improved performance in terms of content delivery and application or service response times.
People |
ORCID iD |
| Ioannis Psaras (Principal Investigator / Fellow) |
Publications
Argyrios G. Tasiopoulos
(2017)
DRENCH: Distributed Load Balancing for NFV-based Service Function Chaining
Argyrios G. Tasiopoulos
(2016)
Tube Streaming: Modelling Collaborative Media Streaming in Urban Railway Networks
Ascigil O
(2017)
On Uncoordinated Service Placement in Edge-Clouds
Ascigil O
(2018)
On-demand routing for scalable name-based forwarding
Ascigil O
(2017)
A keyword-based ICN-IoT platform
Ioannis Psaras
(2016)
Keyword-Based Mobile Application Sharing
Katsaros K
(2016)
Information-Centric Connectivity
in IEEE Communications Magazine
| Description | 2019 - 2020: We are providing a comprehensive overview of our works during the past year, which has expanded to include new areas of research and development, namely work around Blockchains and Distributed Ledger Technology. The text below is also available at: https://www.ee.ucl.ac.uk/~ipsaras/ddec.html The Internet world is turning upside-down. In today's Internet, data is primarily flowing from data-centre servers and server farms, placed largely at the core of the network, towards the users at the edge of the network. In tomorrow's Internet, data will (primarily) be produced at the edge of the network from IoT devices, smart/autonomous vehicles, wearables, sensors and the like. This data will be of enormous volume. It has been said that each autonomous vehicle could generate tens of TBs of data per hour. The current Internet infrastructure is not prepared to accommodate this volume of data coming in from the edge. The current model of sending everything back to the cloud for processing will simply not cope with this wave of data coming from the edge. In order to get the Internet infrastructure prepared for this change, there are a number of components that need to be smoothly integrated into the current Internet architecture. At UCL, we are building solutions to address the needs of a future, privacy-preserving, IoT-dominated edge computing environment. Below we list those necessary components, as well as the works that we have done at UCL over the past few years together with our great colleagues and collaborators. The majority of the solutions and opinions discussed below have been summarised in the following 2-page position paper. Related Publication: Ioannis Psaras, "Decentralised Edge-Computing and IoT through Distributed Trust" (https://www.ee.ucl.ac.uk/~ipsaras/files/edge-distributed-trust.pdf) ACM IoT Open Day@MobiSys 2018, Munich, Germany, June 2018. Store-Process-Send at Edge Data Repositories As a starting point, we have argued that the data communication pattern for edge-computing environments is a "store-process-send" pattern. This is very different to how the Internet is working today. The proposed pattern is building on the assumption that only a small proportion of the data produced at the edge is actually useful. Out of the tens of TBs per hour produced by an autonomous car, quite likely only 1/10th is actually useful information for the car manufacturer, the local council, the (future) car mechanic, or the users themselves. According to the proposed "store-process-send" communication pattern, IoT-produced data is temporarily stored in edge access points; functions then move to the access point (instead of moving data to the function); data gets processed; and, finally, the produced result is sent to its final destination. Mobile Edge Data Repositories We have recently proposed "Mobile Edge Data Repositories" that operate under the "store, process and send" principle. These are edge storage and processing devices to temporarily store incoming data. Think of this storage as a service that the (wireless) ISP is providing its users with, in addition to calls, text, download/upload data. Users have a virtual storage allowance in any access point they connect to. In the following paper we lay out an initial version of the architecture to achieve this goal, based primarily on Information-Centric Networks. Related Publication: Ioannis Psaras, Onur Ascigil, Sergi Rene, Alex Afanasyef, Lixia Zhang, "Mobile Data Repositories at the Edge" (https://www.ee.ucl.ac.uk/~ipsaras/files/edge-data-repositories.pdf) HotEdge Workshop @ USENIX ATC, Boston, USA, July 2018. Computation-Centric Network Architectures Given the huge amounts of data produced at the edge, it has been largely accepted that it is cheaper to bring computation to data, than data to computation. That said, there will soon be a need for a computation-centric network architecture. It turns out that the current model of IP-address based content resolution cannot address "moving functions" that target stored data in edge data repositories, remain active for as long as computation lasts and then dissolve. We have therefore proposed Named Function as a Service (NFaaS), Named Function Mobility (NFM) and Remote Method Invocation (RICE) to address architectural issues for distributed edge computing. Related Publication (also reported last year, but is still very relevant): Michal Krol, Ioannis Psaras, "NFaaS: Named-Functions as a Service" (https://www.ee.ucl.ac.uk/~ipsaras/files/nfaas-icn17.pdf) ACM ICN 2017, Berlin, Germany, September 2017. Related Publication: Michal Krol, Karim Habak, David Oran, Dirk Kutscher, Ioannis Psaras, "RICE: Remote Method Invocation in ICN" (https://www.ee.ucl.ac.uk/~ipsaras/files/rice-remote-method-invocation.pdf) ACM ICN 2018, Boston, USA, September 2018. **Best Paper Award** We have also investigated issues associated to Security in a Named Function Mobility or Named Function Networking environments: Related Publication: Michal Krol, Claudio Marxer, Dennis Grewe, Ioannis Psaras, Christian Tschudin, "Open Security Issues for Edge Named Function Environments" (https://www.ee.ucl.ac.uk/~ipsaras/files/named-function-security-icn-commag-sep18.pdf) IEEE Communications Magazine, 2018. Resource Allocation in Edge-Cloudlets A distributed network of computation spots (or cloudlets) cannot guarantee the elasticity of hugely over-provisioned data centres. In other words, today's assumption of endless resources within data centres does not hold in an edge computing environment. When resources are scarce, resource allocation is becoming a necessary component of the edge-computing system. Assuming that computation spots are spread along the ISP paths, the problem of what computation to host where resembles the caching issues that have been investigated in the past decades. We have therefore, played with existing caching and cache replacement algorithms (i.e., LRU, LFU and the like) to see if they can perform efficiently in case of dynamically instantiated named functions, rather than static content. The results are noteworthy: the performance of well-known replacement algorithms (as well as combinations of them) perform close to optimal. Related Publication (also reported last year, but is still very relevant): Onur Ascigil, Truong Khoa Phan, Argyrios G. Tasiopoulos, Vasilis Sourlas, Ioannis Psaras, George Pavlou, "On Uncoordinated Service Placement in Edge Clouds" (https://www.ee.ucl.ac.uk/~ipsaras/files/cloudcom17-uncoordinated-service-placement.pdf) IEEE CloudCom 2017, Hong Kong, Dec 2017. Market-Based Compute Ecosystem Then, some entity needs to deploy edge computing spots or cloudlets and needs to monetise this infrastructure. We have built several auction-based models for resource allocation, based on demand and supply rules, but also taking into account user mobility. The resulting framework is the foundation of a "Market-Based Compute Ecosystem", run by "In-Networking Computing Providers". The framework includes all the necessary components in order to allocate resources efficiently, on-demand and accommodate mobile users that connect to several Access Points and Base-Stations as they move. Related Publication: Argyrios G. Tasiopoulos, Onur Ascigil, Ioannis Psaras, George Pavlou, "Edge-MAP: Auction Markets for Edge Resource Provisioning" (https://www.ee.ucl.ac.uk/~ipsaras/files/edgeMap-wowmom18.pdf) IEEE WoWMoM 2018, Crete, Greece, June 2018. Related Publication: Argyrios G. Tasiopoulos, Onur Ascigil, Ioannis Psaras, Stavros Toumpis, George Pavlou, "On-path Cloudlet Pricing for Low Latency Application Provisioning" (https://www.ee.ucl.ac.uk/~ipsaras/files/fogspot-lanman18.pdf) IEEE LANMAN 2018, Washington DC, USA, June 2018 - (Invited Paper) Related Publication: Argyrios G. Tasiopoulos, Onur Ascigil, Ioannis Psaras, Stavros Toumpis, George Pavlou, "FogSpot: Spot Pricing for Application Provisioning in Edge/Fog Computing", (https://www.ee.ucl.ac.uk/~ipsaras/files/fogspot-tsc18.pdf) IEEE Transactions on Services Computing (IEEE TSC), January 2019 The Need for Decentralisation Last but not least, a widely distributed computing infrastructure is difficult, if not impossible to manage by a single entity. It is highly unlikely that the edge infrastructure of computation spots will be deployed, run and managed by a single entity/company. Instead, it is more reasonable (and desirable) to assume that a multitude of entities, such as local council, and local authorities together with ISPs, IoT companies and others, will invest to deploy and operate this edge infrastructure. There are a few reasons for that with the most important one being the so-called "War over Data". That is, IoT companies, the automotive industry and any user of the future edge computing infrastructure will (rightly so) do their best to keep user data for themselves. Growing numbers of privacy-preserving technologies, communities and initiatives are building platforms which enable users to keep their personal data for themselves and decide what to share with whom and for what price. Privacy preserving distributed edge computing necessitates Trusted Execution Environments, such as Intel SGX, as well as improved versions of SGX. Such technologies together with blockchain/DLT and smart contract platforms can facilitate processing of data in a secure and privacy-preserving manner, without third parties being able to sniff on personal data, or alter transactions that have already taken place between any two parties. In order to allow for processing of private data on trustless executing nodes and to facilitate a fair and secure customer-provider relationship, we have built Airtnt: a fair payment system for outsourced computation. The protocol was carefully designed to include all the necessary components to avoid any of the involved parties to cheat, cause others to lose their stake, or need a third party to verify computations and payments. We believe Airtnt is a necessary component to provide support for a fundamental feature of the future Internet, that is: decentralisation. Related Publication: Mustafa Al-Bassam, Alberto Sonnino, Michal Król, Ioannis Psaras, "Airtnt: Fair Exchange Payment for Outsourced Secure Enclave Computations", (https://arxiv.org/abs/1805.06411), May 2018. Related Publication: Alberto Sonnino, Michal Król, Argyrios Tasiopoulos and Ioannis Psaras, "AStERISK: Auction-based Shared Economy ResolutIon System for blocKchain", (https://www.ee.ucl.ac.uk/~ipsaras/files/DISS-2019-paper1.pdf) NDSS'19 DISS Workshop, San Diego, USA, February 2017 Related Publication: Michal Król, Alberto Sonnino, Mustafa Al-Bassam, Argyrios G. Tasiopoulos, Ioannis Psaras, "Proof-of-Prestige: A Useful Work Reward System for Unverifiable Tasks", 1st IEEE International Conference on Blockchain and Cryptocurrencies, Seoul, Korea, May 2019. Related Publication: Michal Krol, Ioannis Psaras, "SPOC: Secure Payments for Outsourced Computations" NDSS'18 DISS Workshop, San Diego, USA, February 2018. Other Related Publications C. A. Sarros et al., "Connecting the Edges: A Universal, Mobile-Centric, and Opportunistic Communications Architecture", (https://www.ee.ucl.ac.uk/~ipsaras/files/umobile-commag-connecting-edges.pdf) IEEE Communications Magazine, vol. 56, no. 2, pp. 136-143, Feb. 2018. George Pavlou and Ioannis Psaras, "The troubled journey of QoS: From ATM to content networking, edge-computing and distributed internet governance" (https://www.ee.ucl.ac.uk/~ipsaras/files/qos-comcom.pdf) Elsevier Computer Communications, In Press, Sept 2018 - (Invited Paper) 2018 1) Uncoordinated Service Placement [1] Edge computing has emerged as a new paradigm to bring cloud applications closer to users for increased performance. ISPs have the opportunity to deploy private edge-clouds in their infrastructure to generate additional revenue by providing ultra-low latency applications to local users. We envision a rapid increase in the number of such applications for "edge" networks in the near future with virtual/augmented reality (VR/AR), networked gaming, wearable cognitive assistance, autonomous driving and IoT analytics having already been proposed for edge-clouds instead of the central clouds to improve performance. This raises new challenges as the complexity of the resource allocation problem for multiple services with latency deadlines (i.e., which service to place at which node of the edge-cloud in order to satisfy the latency constraints) becomes significant. The expectation from the mobile edge-/fog-computing paradigm presents, to a certain extend, similarities to the caching era of the 90s. That is, similarly to the move from servers acting as the sole providers of static content to proxy caches and more recently ubiquitous in-network caches (in the Information-Centric Networking area), the edge-/fog-computing paradigm is attempting a shift of computation (as opposed to only static content in the case of caching) closer to the users. By and large, the rationale behind deploying proxy caches was: i) reduce response delay to end-users, ii) reduce core-network traffic, and, iii) reduce server load. Moving to the edge-/fog-computing paradigm, we could realistically argue that the motivation and expectation is roughly similar: move network functions and user-facing applications closer to the users to reduce response delay, network traffic (e.g., in case of heavy data that needs to be uploaded to the cloud) and reduce the ever-increasing stress placed on data centres. In this work, we propose a set of practical, uncoordinated strategies for service placement in edge-clouds, largely inspired by the past work on proxy and in-network content caching. Through extensive simulations using both synthetic and real-world trace data, we demonstrate that uncoordinated strategies can perform comparatively well with the optimal placement solution, which satisfies the maximum amount of user requests. Our results show that there is significant potential to improve the performance of in-network computation by leveraging on the research findings of the past. 2) NFaaS: Named Functions as a Service [2] Recent efforts in the general area of Information-Centric Networking have been focusing on several issues that mainly pertain to traditional content delivery (e.g., routing and forwarding scalability, congestion control and in-network caching). However, in order to keep up with trends in the wider area of future Internet paradigms, there is a pressing need to extend current architectural proposals to support edge-/fog-computing environments. With this goal in mind, we have proposes Named Function as a Service (NFaaS), a framework that extends the Named Data Networking architecture to support in-network function execution. In contrast to existing works, NFaaS builds on very lightweight VMs and allows for dynamic execution of custom code. Functions can be downloaded and be run by any node in the network. Functions can move between nodes according to user demand, making resolution of moving functions a first-class challenge. Our findings demonstrate that a computation-centric, named-function approach to edge-computing can provide significant performance improvements. Requests do not need to be resolved to a central DNS-like system, which is traditionally slow to adapt to network changes (i.e., DNS entries are updated once a day), but instead can be found en-route as the request is travelling towards the cloud. This is a significant finding, given the stringent network latencies required by new and upcoming edge- and fog-computing applications. 3) A Keyword-based ICN-IoT Platform [3] As the Internet of Things (IoT) expands, IoT data should become reusable assets, enabling the creation of new value chains. For example, in a smart city, traffic congestion sensors could be used for both traffic light control and route calculation. Unfortunately, the current end-to-end, silo-based approach to IoT results in the deployment of identical sensors side by side, gathering the same data for different applications and operators. We argue that the edge-computing IoT infrastructure and the data produced by it should be treated as commons. In this work, we have presented the Collaborative Commons Platform for the IoT (Cool-IoT), where data and micro-functions executed at the edge of the network are explicitly named, using software-defined forwarding and lightweight execution virtualisation. Information-Centric Networking (ICN) has been proposed as a promising solution for the Internet of Things (IoT), due to its focus on naming data, rather than endpoints, which can greatly simplify applications. The hierarchical naming of the Named-Data Networking (NDN) architecture can be used to name groups of data values, for example, all temperature sensors in a building. However, the use of a single naming hierarchy for all kinds of different applications is inflexible. Moreover, IoT data are typically retrieved from multiple sources at the same time, allowing applications to aggregate similar information items, something not natively supported by NDN. To this end, in this work we have proposed (a) locating IoT data using (unordered) keywords combined with NDN names and (b) processing multiple such items at the edge of the network with arbitrary functions. We have described and evaluated three different strategies for retrieving data and placing the calculations in the edge IoT network, thus combining connectivity, storage and computing. Our work has the potential to revolutionalise the setup and administration of IoT domains and data management. 4) A Native Content Discovery Mechanism for Information-Centric Networks [4] This work extends the work of entry 5) in 2017's input below, on "Native Content Distribution Network". We have now devised a more complete framework which scales well independently of the number of incoming requests. The results reported in the related paper show significant performance gains. This work has the potential to change the way requests are routed in the network under the Named Data Networking architecture. 5) On the Feasibility of a User-Operated Mobile Content Distribution Network [5] Most of the content that we consume in the Internet today, whether just a text-based webpage, a short video, or a long movie is stored and distributed by Content Distribution Networks (CDNs). CDNs are companies that maintain expensive networks of servers that cover large geographic areas and provide guarantees for low delivery delay. Individual content providers are not able to achieve the same performance, as they cannot afford to maintain such an infrastructure. In order to maintain their specialised infrastructure, CDNs charge high fees. To put this in perspective: if your national TV channel produces a 10min video, which ends up being watched by 1M people, the TV channel is charged ~£3,000. In case of a popular 60min TV series, played once a week and watched by 2M viewers the cost is close to £150,000 per month! CDNs cannot really decrease their charges and still maintain their infrastructure. In order to achieve cost-efficient content distribution we need a radically new approach. We argue that in the era of decreasing hardware costs, it should not be this way. We are building the next generation of CDNs, which works in parallel to traditional CDNs, but does not need large server farms and expensive infrastructure. It needs lots of highly-mobile end- users with smartphone devices. We aim to augment the powerful infrastructure needed to provide timely delivery of content by the users' smartphone devices. In this work, we have explored the potential of a user-operated, smartphone-centric content distribution model for smartphone applications. In particular, we have assumed source nodes that are updated directly from the content provider (e.g., BBC, CNN), whenever updates are available; destination nodes are then directly updated by source nodes in a D2D manner. We leverage on sophisticated information-aware and application-centric connectivity techniques to distribute content between mobile devices in densely-populated urban environments. Our target is to investigate the feasibility of an opportunistic content distribution network in an attempt to achieve widespread distribution of heavy content (e.g., video files) to the majority of the destination nodes. We have proposed ubiCDN as a ubiquitous, user-operated and distributed CDN for mobile applications. Our results have shown that this way of distributing content in the mobile domain can decrease significantly the digital divide, but at the same time reduce the CDN costs of Content Publishers. [1] Onur Ascigil, Truong Khoa Phan, Argyrios G. Tasiopoulos, Vasilis Sourlas, Ioannis Psaras, George Pavlou, "On Uncoordinated Service Placement in Edge Clouds", IEEE CloudCom 2017, Hong Kong, December 2017. [2] Michal Krol, Ioannis Psaras "NFaaS: Named Function as a Service", ACM ICN 2017, Berlin, Germany, September 2017. [3] Onur Ascigil, Sergi Rene, George Xylomenos, Ioannis Psaras, George Pavlou "A Keyword-Based ICN-IoT Platform", ACM ICN 2017, Berlin, Germany, September 2017. [4] Onur Ascigil, Vasilis Sourlas, Ioannis Psaras, George Pavlou "A Native Content Discovery Mechanism for Information-Centric Networks", ACM ICN 2017, Berlin, Germany, September 2017. [5] Ioannis Psaras, Vasilis Sourlas, Denis Shtefan, Sergi Rene, Mayutan Arumaithurai, Dirk Kutscher, George, Pavlou, On the Feasibility of a User-Operated Mobile Content Distribution Network, IEEE WoWMoM 2017 2017 and earlier 1) Distributed Algorithmics for SDN-NFV Environments: We argue for the necessity of an algorithmic resource management framework that captures the involved tradeoffs of NFIs minimum workload, load balancing, and flow path stretch. We have introduced DRENCH [1] as a low complexity NFV and flow steering management framework. Our initial investigation shows that it is indeed possible to design a technically sound architecture to realise distributed content management [2], [3] and service placement [1]. Our algorithmic framework (DRENCH - Distributed Load Balancing for NFV-based Service Function Chaining) [1] provides an elegant and scalable approach to manage the available resources in the network. 2) An Edge-Computing Framework for Smartphone Application Sharing: We have designed the "Keyword-Based Mobile Application Sharing" (KEBAPP) framework [4], according to which smartphone mobile apps are considered as a pooled resource. That is, a mobile user can make use of applications not directly installed on his own mobile device, but can also take advantage of applications installed in nearby devices. The proposed framework has the potential to revolutionalise the way smartphone applications are built and used today, providing a cheap way to realise an edge-cloud infrastructure beyond the boundaries of the fixed Internet. Our work has received attention from the community working on ICT for Development, as our proposed framework is providing a cheap alternative to cloud-like applications. 3) Redefining Mobile Connectivity: We have proposed, designed and built the concept of "Information-Centric Connectivity" (ICCON), a novel concept, which associates connectivity of mobile devices to the content or service they are looking to retrieve [5]. We have argued that link-layer connectivity can be associated with information-availability and in this respect connectivity decisions should be information-aware. Our proposals have the potential to revolutionalise the way mobile applications are resolved, instantiated, and executed. 4) Video Streaming in Mobile Environments: Extending our vision further to the mobile part of the network, we have proposed a model to assess the perceptive Quality of Experience (QoE) of commuters in the London underground network when streaming media [6]. In particular, we have assumed that both WiFi and cellular connectivity exists when trains are within stations and (poor) cellular connectivity exists (i.e., no WiFi) when trains are moving in-between stations. We have found that when users download media individually the perceived QoE is poor. Instead, when users download collaboratively, they achieve significant QoE. Designing smartphone applications taking this concept into account has the potential to enable new services in the mobile part of the network. 5) Towards a Native Content Distribution Network: All current architectural proposals, including the most prevalent one (i.e., NDN) focus on the optimisation of "how to route requests towards the core of the network". We argue that this goes against the original vision of a "native content distribution network" and prevents requests from discovering nearby content, unless the content is on the shortest path to the core of the network. With a view to improving the content discovery capabilities of the NDN architecture, we have proposed an enhancement to the routing fabric, which keeps track of successful (i.e., served) content requests in a separate routing table, called "Downstream FIB" (D-FIB) [7], [8]. The same concept was applied to the case of network fragmentation, where the idea is that once the network gets fragmented and the origin server cannot be reached, the network should intelligently look for copies of the content downwards, towards the users [9]. 6) The placement of content in in-network content caches has a direct impact on the performance of the cache system. Carefully placed content can significantly increase cache hit performance, but at the same time might increase signalling and communication overhead, in order to make informed decisions on both content placement but also content retrieval. We have proposed novel domain clustering techniques in order to limit management overhead but at the same time provide the same level of quality of service (in terms of content delivery time) [3]. Our results demonstrate that indeed, by splitting domains in smaller clusters we manage to limit the overhead required to manage such a cache system without a decrease in cache hits (subject to content popularity). 7) Given the sophistication of network routers under an ICN forwarding and caching paradigm, researchers have investigated the energy efficiency of such a network. There is a tradeoff between caching and therefore, reducing the amount of traffic travelling upwards in the network and the energy needed to perform caching operations. We have implemented a software router and have measured the energy consumption under realistic workloads. We have built a model for the energy consumption of an ICN router, which we hope will be useful for the community and will be used in similar studies [10]. [1] A. G. Tasiopoulos, S. Kulkani, M. Aramathurai, I. Psaras, KK Ramakrishnan, X. Fu, G. Pavlou, "DRENCH: Distributed Load Balancing for NFV-based Service Function Chaining", submitted [2] L. Saino, I. Psaras, G. Pavlou, "Framework and Algorithms for Operator-managed Content Caching", IEEE/ACM Transactions on Networking, submitted [3] V. Sourlas, I. Psaras, L. Saino, G. Pavlou, "Efficient Hash-routing and Domain Clustering Techniques for Information-Centric Networks", Elsevier Computer Networks, 2016. [4] I. Psaras, S. Rene, K. V. Katsaros, V. Sourlas, G. Pavlou, "KEBAPP: Keyword-Based Mobile Application Sharing", ACM MobiArch 2016, Best Paper Award [5] K. V. Katsaros, V. Sourlas, I. Psaras, S. Rene, G. Pavlou, "ICCON: Information-Centric Connectivity", IEEE Communications Magazine, submitted [6] A. G. Tasiopoulos, I. Psaras, V. Sourlas, G. Pavlou, "Tube Streaming: Modelling Collaborative Media Streaming in Urban Railway Networks", IFIP Networking 2016. [7] V. Sourlas, O. Ascigil, I. Psaras, G. Pavlou, "Opportunistic Off-Path Content Discovery in Information-Centric Networks", IEEE LANMAN 2016, Best Paper Award. [8] V. Sourlas, O. Ascigil, I. Psaras, G. Pavlou, "Enhancing Information Resilience in Disruptive Information-Centric Networks, IEEE Transactions On Network and Service Management (IEEE TNSM), submitted. [9] Vasilis Sourlas, Leandros Tassiulas, Ioannis Psaras and George Pavlou, "Information Resilience through User-Assisted Caching in Disruptive Content-Centric Networks" 14th IFIP NETWORKING, Toulouse, France, May 2015. Best Paper Award. [10] K. Ohsugi, J. Takemasa, Y. Koizumi, T. Hasegawa, I. Psaras, "Power Consumption Model of NDN-based Multicore Software Router based on Detailed Protocol Analysis", IEEE Journal on Selected Areas in Communications, 2016. |
| Exploitation Route | The outcomes of the project are mainly of academic interest, but not exclusively. The ultimate purpose of our contributions is to increase the visibility of our group and attract interest from academic and industry partners. The findings and achievements, e.g., our Best Paper Award, is already contributing towards this direction with invitations for talks and invitations to join research grant proposal consortia. We have received active interest from private institutions working on Edge Computing to further investigate and potentially apply the findings of our work in the Automotive Sector. The Automotive Edge Computing Consortium (AECC - https://aecc.org/) is actively looking to address issues related to storage and computation of huge volumes of data at the edge of the network. Our work outlined above is addressing those issues directly and our work has received attention and visibility from stakeholders in this area. Our model in [6] above has already attracted attention (through talks where we have already disseminated our work) by the community and we expect that after publication it will be established as one of the main models to assess the QoE perceived by mobile users in transport networks. Our findings in the NFaaS work demonstrate that a computation-centric, named-function approach to edge-computing can provide significant performance improvements. Our work on Keyword-based IoT has the potential to revolutionalise the setup and administration of IoT domains and data management. Our results on ubiCDN have shown that this way of distributing content in the mobile domain can decrease significantly the digital divide, but at the same time reduce the CDN costs of Content Publishers. Our work on dynamic service placement [1] is one of the very few of its kind in the area and as such, we expect that it will be widely adopted and used in the areas of SDN and NFV. Finally, we expect that our work on "Keyword-Based Mobile Application Sharing" and "Information-Centric Connectivity" will spark a new research direction and will trigger further research, which will build on the concepts proposed in this project. In fact, KEBAPP has already attracted attention by the research community working on ICT for Development, as it provides a very cheap way to provide cloud-like services at the edge of the network. Indeed, community networks in developing countries can greatly benefit by such a framework. We have established channels with the ICT for Development community and we plan to disseminate our findings further. |
| Sectors | Communities and Social Services/Policy,Creative Economy,Digital/Communication/Information Technologies (including Software),Transport |
| URL | https://www.ee.ucl.ac.uk/~ipsaras/ddec.html |
| Description | Our work on Distributed and Decentralised Edge Computing (DDEC@UCL - https://www.ee.ucl.ac.uk/~ipsaras/ddec.html) is providing a complete and mature framework for the challenges faced by the Automotive sector. We have received active interest from private institutions working on Edge Computing to further investigate and potentially apply the findings of our work in the Automotive Sector. The Automotive Edge Computing Consortium (AECC - https://aecc.org/) is actively looking to address issues related to storage and computation of huge volumes of data at the edge of the network. Our work outlined above is addressing those issues directly and our work has received attention and visibility from stakeholders in this area. There is recently a trend towards Decentralisation of Internet Services, mainly due to advances in the area of Blockchains and Distributed Ledger Technologies. The trend is mainly led by non-profit organisations that manage open-source projects. Our contributions in this area, as detailed in the Key Findings section and the Publications section are already making an impact in this area and are considered for adoption by some of the leading foundations and companies in the area. In particular, our work on "Proof-of-Prestige: A Useful Work Reward System for Unverifiable Tasks" is providing an elegant solution for securing permissionless networks by requiring users to stake part of their wallet balance (monetary value) in order to participate in the network. When users misbehave, they are penalised through the stake that they have put in the system. Our unique method of securing permissionless networks has already attracted attention and a modified version of the concept is currently being integrated in one of the leading cryptocurrencies in the area of decentralised storage. We believe that as the area and the field matures, our works will be among the leading forces in developing the new decentralised Internet landscape. In 2020 our work on Proof of Prestige has been picked and the concept has been integrated in the protocol design of the Filecoin Blockchain, the first of its kind decentralised storage network with its own blockchain. The work has motivated part of the thinking in the design of the message propagation protocol for blocks and transaction messages. Furthermore, in 2021, and in the same line of work, the company that supports the development of Filecoin is building a decentralised content delivery network (CDN). In the design phase of this project, several of our works, including the "Proof of Prestige" mentioned above, but also extending to other papers on edge computing (e.g., "Edge-MAP") have played a central role in setting the design foundations of this new type of networks. This is an extremely important development as decentralised storage and delivery networks are expected to become the new emerging topic in large-scale computer networks. In the period 2020-2021, the PI has provided Knowledge Transfer services to leading companies in the area, such as Protocol Labs and the Ethereum Foundation. The PI has received an offer and has spent 8 months to date in sabbatical advising Protocol Labs as a Research Scientist. At the same time, our work on edge computing and information-centric networks has been recognised by major players in the area, i.e., Cisco and our team has received a research grant in the form of gift funding for ~15 months. A part of our work in the INSP project has been promoted and is currently being supported by initiatives in the area of ICT for development. In particular, the ISOC-supported (Internet Society) workshop on "Future Internet for Development" (FI4D - https://www.cl.cam.ac.uk/~al773/fi4d/2017/, co-located with IEEE Consumer Communications and Networking Conference (CCNC)) produced a report on how our findings on "Keyword-Based Mobile Application Sharing" (reported in more detail in the "Key Findings" section) can be of use in developing regions. Our contribution (the software tools developed), which can be part of a cheap hardware infrastructure (such as a Raspberry Pi) can provide edge-cloud services to the general public in developing countries. Although there is no tangible outcome or deployment yet, the fact that our work is supported by ISOC (apart from the research/academic community) is a positive sign for future societal impact. The report of the workshop can be found here: https://www.ee.ucl.ac.uk/~ipsaras/files/fi4d-report-final.pdf |
| First Year Of Impact | 2016 |
| Sector | Communities and Social Services/Policy,Digital/Communication/Information Technologies (including Software),Financial Services, and Management Consultancy,Transport |
| Impact Types | Cultural,Societal,Economic |
| Description | Cisco Gift on Hybrid ICN |
| Amount | $120,000 (USD) |
| Organisation | CISCO Systems |
| Sector | Private |
| Country | United States |
| Start | 07/2020 |
| End | 10/2021 |
| Description | EC H2020 - ICN2020 |
| Amount | € 1,300,000 (EUR) |
| Funding ID | Grant Agreement No. 723014 |
| Organisation | European Commission |
| Sector | Public |
| Country | European Union (EU) |
| Start | 07/2016 |
| End | 05/2019 |
| Description | EU H2020 DECODE |
| Amount | € 100,000 (EUR) |
| Organisation | European Commission H2020 |
| Sector | Public |
| Country | Belgium |
| Start | 04/2019 |
| End | 12/2019 |
| Description | UCL EEE Dept Internal Grant for Laboratory and Equipment Support |
| Amount | £25,000 (GBP) |
| Organisation | University College London |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 05/2017 |
| End | 04/2022 |
| Description | Research Advisor to Protocol Labs Inc. |
| Organisation | Protocol Labs Inc. |
| Country | United States |
| Sector | Private |
| PI Contribution | I have advised Protocol Labs (through a formal consultancy agreement) in issues related to architecture of the network that the company is building. Protocol Labs is building a decentralised, peer-to-peer, content-addressable network. The approach used by Protocol Labs is very similar to several research works carried out by my team and therefore, the expertise has been very valuable to the ongoing activities within Protocol Labs. |
| Collaborator Contribution | Protocol Labs has contributed in this collaboration by allocating members of their team to our endeavours in diving deeper into research questions. |
| Impact | I have helped Protocol Labs crystalise their approach to optimising content routing, timely content delivery, but also formally define a series of open problems. I have helped build a new research lab and identify its structure, goals and research directions. The PI has been offered a longer-term position as a Research Scientist at Protocol Labs to lead several projects. The PI has subsequently got a temporary sabbatical from his position at UCL. So, in summary, the PI has served as a research advisor to Protocol Labs from June 2019 to July 2020 and then as a Research Scientist (on sabbatical from UCL) from August 2020 to the present day. |
| Start Year | 2019 |
| Title | AStERISK: Auction-based Shared Economy ResolutIon System for blocKchain |
| Description | Recent developments in blockchains and edge computing allows to deploy decentralized shared economy with utility tokens, where altcoins secure and reward useful work. However, the majority of the systems being developed, does not provide mechanisms to pair workers and clients, or rely on manual and insecure resolution. AStERISK bridges this gap allowing to perform sealed-bid auctions on blockchains, automatically determine the most optimal price for services, and assign clients to the most suitable workers. AStERISK allows workers to specify a minimal price for their work, and hide submitted bids as well the identity of the bidders without relying on any centralized party at any point. We provide a smart contract implementation of AStERISK and show how to deploy it within the Filecoin network, and perform an initial benchmark on Chainspace. |
| Type Of Technology | Software |
| Year Produced | 2019 |
| Open Source License? | Yes |
| Impact | This is a useful tool to assign users to service providers in a decentralised storage network environment. It is already been tested in a commercial decentralised storage network environment and considered for adoption by the related company. |
| URL | https://www.ndss-symposium.org/ndss-paper/auto-draft/ |
| Title | Airtnt: Fair Exchange Payment for Outsourced Secure Enclave Computations |
| Description | We present Airtnt, a novel scheme that enables users with CPUs that support Trusted Execution Environments (TEEs) and remote attestation to rent out computing time on secure enclaves to untrusted users. Airtnt makes use of the attestation capabilities of TEEs and smart contracts on distributed ledgers to guarantee the fair exchange of the payment and the result of an execution. Airtnt makes use of off-chain payment channels to allow requesters to pay executing nodes for intermediate "snapshots" of the state of an execution. Effectively, this step-by-step "compute-payment" cycle realises untrusted pay-as-you-go micropayments for computation. Neither the requester nor the executing node can walk away and incur monetary loss to the other party. This also allows requesters to continue executions on other executing nodes if the original executing node becomes unavailable or goes offline. |
| Type Of Technology | Software |
| Year Produced | 2018 |
| Open Source License? | Yes |
| Impact | The solution proposed here is important in the area of decentralised cloud computing. It enables anyone with space CPU capacity to fairly trade CPU power for monetary reward (in the form of cryptocurrency). The software is already attracting attention from industrial players in the blockchain space. |
| URL | https://arxiv.org/abs/1805.06411 |
| Title | Computation Offloading in Information-Centric Networks |
| Description | This software package implements a computation-centric architecture over Named-Data Networks (NDN). The system is able to perform in-network load balancing of incoming computation requests, reliably authenticate consumers and allow them to submit large payloads without routable prefixes. The system is able to migrate requested functions in the form of unikernels where they are requested and allocated, follows the Information-Centric Networking pull-based model and introduces only minimal changes to the NDN stack. |
| Type Of Technology | Software |
| Year Produced | 2019 |
| Open Source License? | Yes |
| Impact | This system is the first of its kind to be developed over an ICN architecture and has already received wide attention by the community with several active users. The related paper that has been published alongside it has received the Best Paper Award in the ACM Conference on Information-Centric Networks, the flagship event in the area. The link to the paper itself is here: https://conferences.sigcomm.org/acm-icn/2018/proceedings/icn18-final9.pdf and the links to the demo and the code are given below. |
| URL | https://conferences.sigcomm.org/acm-icn/2018/proceedings/icn18posterdemo-final4.pdf |
| Title | KEBAPP |
| Description | In the context of our work on Keyword-Based Mobile Application Sharing (KEBAPP - see Key findings for further details), we have developed a software platform based on innovative Information-Centric Networking technology. Our platform enables mobile devices (e.g., smartphones, tablets) to share applications with nearby devices/users. As a representative example, we have developed the "routeFinder" application. Picture the following: you have just arrived at a busy train station or airport in a foreign country where you have to pay roaming charges to access the Internet. You need to find the route to your hotel/meeting through public transport, but don't have the local transport-authority application (e.g., similar to TfL). Chances are that the majority of people around you have the application you need. Using our novel platform, you can make a request giving free text keywords (e.g., your journey details) and devices around will calculate the route (e.g., which buses, tube lines you need to use) and return back the result. This is of great use not only to avoid roaming charges, but also avoid slow Internet connections when in very crowded places, or places where Internet is not available (e.g., while inside the tube, airplane etc.). |
| Type Of Technology | Software |
| Year Produced | 2016 |
| Impact | The KEBAPP software has the potential to revolutionise the way smartphone applications are built. We're still in the early stages of development so there are no further impact beyond academic publication/dissemination. In the future, we plan to make the software open-source to allow the community to build on top of it and create new applications. |
| URL | http://dl.acm.org/citation.cfm?doid=2980137.2980141 |
| Title | Proof of Prestige: A Useful Work Reward System for Unverifiable Tasks |
| Description | As cryptographic tokens and altcoins are increasingly being built to serve as utility tokens, the notion of useful work consensus protocols, as opposed to number-crunching PoW consensus, is becoming ever more important. In such contexts, users get rewards from the network after they have carried out some specific task useful for the network. While in some cases the proof of some utility or service can be proved, the majority of tasks are impossible to verify. In order to deal with such cases, we design "Proof-of-Prestige" (PoP)-a reward system that can run on top of Proof-of-Stake blockchains. PoP introduces "prestige"which is a volatile resource and, in contrast to coins, regenerates over time. Prestige can be gained by performing useful work, spent when benefiting from services and directly translates to users minting power. PoP is resistant against Sybil and Collude attacks and can be used to reward workers for completing unverifiable tasks, while keeping the system free for the end-users. We use two exemplar use-cases to showcase the usefulness of PoP and we build a simulator to assess the cryptoeconomic behaviour of the system in terms of prestige transfer between nodes. |
| Type Of Technology | Software |
| Year Produced | 2018 |
| Impact | Proof of Prestige is an improved version of Proof of Stake algorithms, which will prove useful in many different blockchain-based platforms and applications. The paper has been accepted in the 1st International Conference on Blockchain and Cryptocurrencies (http://icbc2019.ieee-icbc.org/). The code is released as open-source and the project is disseminated in order to attract visibility and build a user-base around it. |
| Description | Internet Engineering Task Force (IETF) and Internet Research Task Force - Decentralised Internet Infrastructure (DINRG) |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Industry/Business |
| Results and Impact | Dr Michal Krol, member of our group working on the INSP Fellowship project and Dr Ioannis Psaras (PI) have actively participated and presented our recent work on the newly-formed DINRG group. The group focuses on issues related to the decentralisation of the Internet Infrastructure. |
| Year(s) Of Engagement Activity | 2018,2019,2020 |
| URL | https://datatracker.ietf.org/meeting/interim-2018-dinrg-01/materials/minutes-interim-2018-dinrg-01-2... |
| Description | Internet Engineering Task Force (IETF) and Internet Research Task Force - Information-Centric Networking Research Group (ICNRG) |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Industry/Business |
| Results and Impact | The Internet Research Task Force (IRTF) is the research "arm" of the Internet Engineering Task Force (IETF), the main standardisation body of Internet protocols and related activities. The Information-Centric Networking Research Group (ICNRG) has been the main group promoting activities related to the ICN paradigm. The PI is actively engaging with the group since its early days. He has contributed to one of the first RFCs of the group "RFC 7927: ICN Research Challenges". The group is meeting four times a year. We have been active in most of the meetings of the group and have presented our results which have triggered extensive discussions in the community. This way, we promoted our work and created visibility around our group and the activities in the project. We will be hosting the next interim meeting of the group during the London IETF in March 2018. The interim meeting of the ICRNG will take place at UCL. |
| Year(s) Of Engagement Activity | 2016,2017,2018 |
| URL | https://trac.ietf.org/trac/irtf/wiki/icnrg |
| Description | RFC 7927 - Internet Engineering Task Force (IETF) and Internet Research Task Force - Information-Centric Networking Research Group (ICNRG) |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | The Internet Research Task Force (IRTF) is the research "arm" of the Internet Engineering Task Force (IETF), the main standardisation body of Internet protocols and related activities. The Information-Centric Networking Research Group (ICNRG) has been the main group promoting activities related to the ICN paradigm. The PI is actively engaging with the group since its early days. He has contributed to one of the first RFCs of the group "RFC 7927: ICN Research Challenges" and is regularly presenting the progress of this project to the ICNRG group. |
| Year(s) Of Engagement Activity | 2016,2017,2018,2019,2020 |
| URL | https://trac.tools.ietf.org/html/rfc7927 |