NDN Community Meeting 2025 (NDNComm 2025)

Abstract:

We present Ownly - a secure decentralized workspace application built with Named Data Networking and Security. In this talk we will deep dive into the network internals, the security model and details such as Ownly’s usage of NDN Testbed infrastructure.

Abstract:

Modern scientific workloads can be time-consuming and multi-step porcesses. For instance, the workflows in genomics research involve managing large volumes of data, computing resources, and multiple steps in a distributed environment. We propose a federated solution, SASW, that can utilize available resources across distributed facilities without any central controller. SASW converts workflows into Directed Acyclic Graphs (DAGs) of tasks, and available resource nodes self-orchestrate parallel executions of independent tasks based on the data availability and shared system state. SASW is also designed with a distributed logging system for task auditing, fine-tuning performance, and workflow debugging.

Abstract:

Solid, initially developed by Sir Tim Berners-Lee’s Lab at MIT and their collaborators, allows people to manage and control personal data using Personal Online Data Stores (Pods). The technology ensures that data is no longer centralised around applications but instead is centred around the user. This shift enables individuals to decide how their data is collected, accessed, stored, and shared, addressing growing concerns about privacy, transparency, and control in the digital world.

Abstract:

A network protocol architecture provides the substrate on which applications are built. This paper uses the development of a specific application, a user-controlled collaborative editor (UCE), as a case study to explore its realizations on two different network protocol architectures, TCP/IP and Named Data Networking (NDN), and to understand the impact of the architectural differences on the realization of UCE. We found that the multiple Internet enhancements and entrenchments over the years created technical barriers for direct user-to-user communications which are needed in realizing IP-based UCE (IUCE), and consequently resulting in layers of complexity and inefficiency. We further noticed that these layers of function indirections roughly approximate the functions that NDN natively provides to application development. This case study demonstrates that a network protocol architecture can bring profound impacts on application development and complexity, and hence the value of reopening the exploration of new Internet architecture designs.

Abstract:

AI data requirements have been thoroughly established in the context of training. AI Inferencing is less understood with a much more fragmented approach. NDN has inherent strengths to assist in the establishment of capabilities for scaling AI inference.

Abstract:

This talk discusses the opportunities for named data in emerging XR applications through the lens of a specific production currently underway at UCLA—an experimental, immersive version of the Broadway musical Xanadu that incorporates on-premises extended reality, cloud-hosted generative AI, and phone-based audience interaction. We will take a brief tour of namespaces and named-based interactions implemented across various IP-based services—Firebase, Amazon Web Services, Unreal Engine, etc.—as well as application-layer protocols. Through these specific examples, the talk will explore two questions: What would be better if they used NDN? And, why don’t they?

Abstract:

The interenet has been under relentless scaling pressure, and it is a testament to the robustness of the Internet’s basic datagram architecture that we’ve been able to scale the network by a billion-fold and still retain most of the Internet’s basic behaviours. However there has been a fundamental shift in this environment and that is the move in the locus of attention (and investment) up the protocol stack from transmission and networking to the application layer. The result is that these days the proponderance of content and services are delivered from replicated content delivery platforms, and the basic means of traffic switching has shifted from routing protocols to the DNS anmd the applcation. We will look at why this has occurred and how this related to the concepts of named data networking.

(the form will not permit me to upload a pdf - the pdf can be found at https://www.potaroo.net/presentations/2025-04-17-ndncomm.pdf)

Abstract:

NDNd is a Golang implementation of the Named Data Networking stack. In this talk, we will discuss the NDNd structure and design decisions and focus on usability, the current status of implementation, and some results including performance benchmarks.

Abstract:

ndn-dv is a new distance-vector routing protocol for Named Data Networks. The ndn-dv design minimizes routing overhead by separating prefix and router reachability in the NDN network. We will discuss the design and implementation of ndn-dv along with some evaulation results.

Abstract:

The high frequency of connectivity changes in MANET scenarios makes running a regular network routing protocol infeasible. This talk reports our investigation into using NDN forwarding strategies alone in Interest forwarding in a bandwidth-constrained MANET scenario. We used Mini-NDN to conduct various experiments. We show that the baseline ASF strategy, which is designed for the infrastructure environment, exhibits high traffic overhead, which causes wireless network congestion. We modified ASF and show the new version is able to perform well in the examined test cases.

Abstract:

Defined-trust Domains (DtD) are a type of Limited Domain [RFC8799] where the rules specifying the networking of application information are defined in a communications schema that governs the information exchanged and is integrated into the DtD’s Defined-trust Transport Protocol (DeftT). The schema also defines the form of valid member identities which are distributed as a chain of trust rooted at the trust anchor of the DtD. This framework grew out of work with NDN [DNMP] and background can be found at [IOTK,TST,DTLD]. Example implementations [DCT]. With a focus on secured communications, DtDs are constructed to mitigate attacks, including replay.

DtD applications interface to collections of hierarchically named publications. Collections are synchronized across all members of the DtD by a sync protocol that sends and receives the collection’s packets wrapped in its own protocol data units (PDUs) which are exchanged using system transports, e.g., UDP, TCP, IPv6. PDUs are hierarchically named and prefixed with their collection name. Publications are timestamped which helps to prevent replay attacks, among other uses. In deployment, Scott Gray of Operant Networks found devices that do not use NTP and are sufficiently different that publications were being discarded. For timestamps to be meaningful, the members of a DtD must use a (virtual) clock that is synchronized throughout the DtD. The virtual clock does not need to be monotonic or particularly close to actual time and the relative precision of this clock in a common DtD can be  0.5 seconds (for publication lifetimes   1 second), though the aim is to keep it well below that value.

Clock synchronization has many solutions: NTP and its descendants for synchronizing internet-connected systems and a number of approaches for wireless sensor networks that aim for microsecond accuracy so carefully account for processing delays and make use of physical or MAC layer clocks. Both these classes of approach are focused on synchronizing the system clock of their devices. DtDs need a clock that can be used throughout the Domain for publications. Its granularity can be closely related to its publication lifetimes and it need not be monotonic nor relate to any global clock (e.g., GMT). This talk describes a Trust Domain Virtual Clock that is integrated with DeftT and completely self-contained [TDVC]. TDVC contains a delay estimator that can be used for DtD subnets with transit delays that exceed calibration granularity.

TDVC has its own DeftT collection that propagates virtual clock publications via single-hop reachable members (neighbors), gossip-style, to minimize clock noise due to transmission delay. Contributions to the calculation are weighted by the neighborhood size of each member; that is, a member with more neighbors is given more weight than a member with fewer neighbors. Members use the differences between their virtual clock and that of its neighbors in calibration computations. A member considers its virtual clock to be calibrated when all neighbor difference values are within a tolerance (that must account for the expected transmission delay) and all of its neighbors also indicate they are in tolerance. Calibration restarts either when a neighbor’s clock is detected out of tolerance or a neighbor sends values from a new calibration. In DtDs that contain relay elements, the relays act like a meshed member that connects otherwise disconnected members. The talk will cover TDVC details and results from initial testing.

[DNMP] Nichols, K., "Lessons Learned Building a Secure Network Measurement Framework Using Basic NDN", Proceedings of ACM ICN ’19, September 24-26, 2019, Macao, China. [IOTK] Nichols, K., "Trust schemas and ICN: key to secure home IoT", Proceedings of ACM ICN ’21, September 22-24, Paris, France, https://doi.org/10.1145/3460417.3482972. [TST] K. Nichols, "Versec/DCT to create and use trust schemas", ACM ICN 2021 Tutorial, https://conferences2.sigcomm.org/acm-icn/2021/assets/tutorial-trust-schema/2-versec-dct- e1c6ddae2a10c47df55846b58ab84b93aeaff30d1eb00fa4fe9383afb3abca59.pdf [DTLD] Nichols, K., Jacobson, V., King, R., "Defined-trust Transport for Limited Domains”, https://pollere.net/Txtdocs/defTrust.html [TDVC] Nichols, K. “Defined-trust Domain Virtual Clock”, https://www.tdcommons.org/dpubs_series/7887