Oct 8, 2020 — Introduction to the ICN-WEN Program and Learnings from NDN will give a brief overview of the setup of the center.
143 KB – 16 Pages
PAGE – 1 ============
Session 1: NDN Trial Deployments Publishing Genomics Datasets into NDN Testbed and Integrating with Cloud Wows F. Alex Feltus Clemson University Susmit Shannigrahi Tennessee Tech University Stephen Ficklin Washington State University Rini Pauly Clemson University Cole Mcknight Clemson University David Reddick Tennessee Tech University Tyler Biggs Washington State University Cameron Ogle Clemson University Background The genomics community has made astronomical progress in recent decades. Thanks to medium-cost DNA sequencing machines, the genomics community is rapidly approaching the computational petascale at universities and research centers. For example, in 12 years the SRA repository at the National Center for Biotechnology Information (NCBI) has accumulated >42 petabytes of high-throughput DNA sequence data. There are many other similar genomic data repositories around the world including Japan’s DDBJ, EU/UK’s Ensembl, and NASA’s GeneLab. All datasets are complemented with valuable metadata representing evo -lutionary relationships, biological sample sources, measurement techniques, and biological conditions. While there is metadata overlap between repositories, a common data aggregation site for sharing data is needed | an NDN framework. Furthermore, it is common for users of compute resources in close proximity to use the same genomic datasets. By using the NDN framework with its built-in caching mechanism, commonly used datasets can exist closer to where they are being analyzed. The result is a framework for unprecedented data aggregation between heterogeneous metadata environment and accessibility that can be applied to many disciplines across computational science. Named Genomics Data As part of the NSF SciDAS project (#1659300), we have developed Pynome, a Python com-mand line interface tool that provides the user with a way to download desired genome as-sembly from the Ensembl database (https://github.com/SystemsGenetics/pynome). We have preprocessed all multicellular genomes and given them names based upon standardized evo -lutionary relationship metadata as well as repository metadata. Our data naming structure is semantically meaningful, hierarchical, and will make sense to the bioinformaticists across ge-nomics communities. Here is an example data naming hierarchy for an NCBI SRA gene ex-pression dataset: /BIOLOGY/SRA/9605/9606/NaN/RNA-Seq/ILLUMINA/TRANSCRIPTOMIC/PAIRED/Kidney/ PRJNA359795/SRP095950/SRX2458154/SRR5139394/SRR5139394_1>SRR5139394_1.fastq.gz 1
PAGE – 2 ============
These genomes are being published into the NDN scien testbed that was created at Colorado State and since been extended to sites across the country. We are also publishing gene expression datasets from the 42+ PB SRA repository that includes NCBI-SRA-Animal Example (0.16 TB; 36X2 SRA Samples; Human kidney) and NCBI-SRA-Plant Example (1.14 TB; 475X2 SRA Samples; Rice leaves). These datasets were published using NDN-Python-repo and available to genomics researchers for download. Ww-NDN Testbed Interaction We have containerized all NDN software and tested those containers in the Google Cloud Platform (GCP), P Research Platform (PRP), and TACC Rodeo Kubernetes (K8s) clusters. We have pulled data from the NDN testbed into genomics wws using the Data Transfer Pod (DTP; https://github. com/SciDAS/dtp) that can pull data from multiple sources and land the data onto a K8s namespace PVC. Finally, we have adapted the GEMmaker https://github.com/SystemsGenetics/GEMmaker) containerized ww to run w gene expression mapping containers on each K8s system. Advantages Publishing datasets over NDN allows for popular datasets to be automatically cached in the network. The datasets can be cached in the cloud platform running ww containers, signtly speeding up data retrieval. Additionally, the researchers can deploy jobs to cloud locations where data is already retrieved and available, reducing time for completing a ww. Linking DTP with NDN is an improvement over the current methods where the containers must be with the URLs of the datasets. Once deployed, the containers stick to the sources of data. With NDN, the data source can change based on network conditions, data availability, and the distance between computation and data. Since NDN provides a more way to cache and distribute data to Genomics wws, we have also created an HTTP-NDN interface. When requested data is not available in the NDN testbed, we automatically download and publish the datasets into the NDN testbed using the naming scheme we describe above. The b of retrieving data over NDN are location independence, in-network caching, and forwarding of requests. Data-Centric Ecosystems for Large-scale Data-Intensive Science Edmund Yeh Northeastern University Data-centric networking is becoming increasingly important for large-scale data-intensive science. This talk will discuss recent results and future prospects for building data-centric distribution, caching, access, and analysis systems for major science programs. We discuss the NSF-funded SANDIE project, a collaboration among Northeastern, Caltech and Colorado State Universities, which aims to develop and deploy a Named Data Networking (NDN)-based platform for the Large Hadron Collider (LHC) high energy physics program, one of the world’s largest big data applications. This project recently a major demo at SC19, which showed live that NDN can tly index and deliver LHC high energy physics data over a transcontinental layer-2 testbed (Boston-Denver-Los Angeles) at over 6.7 Gbps (using a single thread). The demo also showed how optimized caching algorithms can decrease download times by a factor of 10. We next introduce the newly awarded NSF project N-DISE, led by Northeastern, Caltech, UCLA, and Tennessee Tech, which will produce a highly t and NDN-based petascale data distribution, caching, access, and analysis system serving major science programs including LHC high energy physics and the BioGenome and human genome projects. Building on the results of SANDIE, N-DISE will develop high-throughput caching and forwarding methods, containerization techniques, hierarchical memory management subsystems, congestion control mechanisms, integrated with FPGA acceleration subsystems, to produce a system capable of delivering LHC and genomic data over wide area networks at throughputs approaching 100 Gbps, with tly decreased download times. 2
PAGE – 3 ============
mGuard Project Overview Lan Wang Santosh Kumar Lixia Zhang University of Memphis University of Memphis UCLA This talk will give an overview of the new NSF funded mGuard project. mGuard aims to address two major data access challenges encountered by the NIH Center of Excellence for Mobile Sensor Data-to-Knowledge (MD2K) in its pursuit to share mobile health (mHealth) data among researchers who investigate a wide range of health and wellness issues. First, because wearable sensor data may expose privacy-sensitive information about a user, they should be accessed only by authorized users; currently this access control is largely handled manually, incurring high overhead and subject to human errors. Second, to enable real-time intervention for certain medical conditions, researchers need to retrieve and process the sensor data in real-time, which is not supported at this time. mGuard tackles the above challenges by utilizing the results from the NSF-supported Named Data Networking (NDN) initiative, in particular the solutions that automate the cryptographic key management for data access control (name-based access control, or NAC) and the solutions that enable real-time synchronization among distributed datasets (NDN Sync). First, mGuard utilizes and extends NDN NAC to automate access control of tial data to authorized researchers. Second, it utilizes NDN Sync to provide real-time data production based on this, it enables applications to publish and subscribe to data in real time by directly using MD2K data names. These new capabilities will be deployed in the MD2K cyberinfrastructure. FABRIC -Capabilities and Use-cases Ilya Baldin RENCI/UNC Chapel Hill FABRIC is a 4-year NSF-funded Mid-Scale construction project that by 2023 will deploy a world-wide experimental network infrastructure intended to support experimentation with stateful network protocols, applications and architectures. Using a combination of dedicated 100G and Terabit links the infrastructure will link FABRIC nodes together with experimental testbeds (NSF Clouds, PAWR facilities), HPC centers, scien instruments and public clouds to create a rich environment for experimentation with a variety of applications and protocols that want to take advantage of in-network data processing, fusion and storage. This talk will describe the status of FABRIC project and its expected capabilities in supporting this advanced experimentation Session 2: ICN for Wireless Edge Networking Introduction to the ICN-WEN Program and Learnings from NDN Srikathyayani Srikanteswara Intel The joint Intel/NSF ICN-WEN Center recently completed its 3 year tenure successfully. This short talk will give a brief overview of the setup of the center. The talk will also touch upon some of our own experiences with NDN during this journey, and ways to increase its value proposition for industry. SPLICE: Secure Predictive Low-Latency Information Centric Edge for Next Generation Wireless Networks Srinivas Shakkottai Texas A&M University The generation of wireless communication promises to provide Gigabits per second data transfer rates 3
PAGE – 4 ============
and communication delays of less than a millisecond. t challenges must be overcome in designing a system architecture such that data intensive and/or latency sensitive applications can obtain the information that they need for peak performance. Information-Centric Networking (ICN) has the potential to enable wireless that are critical to support the strict guarantees desired by new applications such as virtual and augmented reality (VR/AR). The goal of this project is to design, develop and demonstrate SPLICE, a Secure Predictive Low-Latency Information-Centric Edge wireless network that will be able to provide information guarantees to such emerging applications. This talk will discuss the major results of the project and the way forward. The project is a collaborative across Texas A&M University, the Ohio State University, Purdue University, Washington University at St. Louis, and the University of Illinois at Urbana-Champaign. Update on ICN-Enabled Secure Edge Networking with Augmented Re-ality (ICE-AR) Burke UCLA REMAP This short talk will provide an overview and update on the ICN-Enabled Secure Edge Networking with Augmented Reality (ICE-AR) project, which is supported by the NSF/Intel ICN-WEN program. ICE-AR is a collaborative of UCLA, Florida International University, and New Mexico State University. Finishing its third year, the project explores how Named Data Networking can support next generation wireless edge networking applications by integrating cross-layer wireless optimizations, data-centric security, acceleration-as-a-service, and new, ICN-inspired application concepts. This talk will provide a brief project overview and highlights from the last year, along with next steps for the work. Session 3: Routing and Forwarding On the Granularity Problem in NDN Adaptive Forwarding Teng Liang Peng Cheng Laboratory Junxiao Shi National Institute of Standards and Technology Beichuan Zhang The University of Arizona One unique architectural b of Named Data Networking (NDN) is adaptive forwarding, i.e., the forwarding plane is able to observe data retrieval performance of past Interests and use it to adjust forwarding decisions of future Interests. To be e, adaptive forwarding assumes what we call Interest Routing Locality, that Interests sharing the same are likely to take the same or similar forwarding path within a short time window, thus past observation can be an indicator of future performance. Since Interests can have multiple common with t lengths, the real challenge is what length should be used in adaptive forwarding. The longer the common is, the better Interest Path Locality, but the fewer future Interests it covers and the larger the forwarding table size. Existing implementations use static length, which is known to have problems in dealing with partial network failures. In this presentation, I will introduce our work on dynamically aggregate and de-aggregate name in the forwarding table, so to use the that are the most appropriate under the current network situation. To reduce the overhead, we design mechanisms to minimize the use of longest match in the processing of Data packets. Simulations demonstrate that the proposed techniques can make better forwarding decisions under partial network failures with cantly reduced overhead. 4
PAGE – 5 ============
m-ASF -An Adaptive SRTT-based Forwarding Strategy for Mobile En-vironments Muktadir Chowdhury University of Memphis Alexander Lane University of Memphis Lan Wang University of Memphis Adaptive SRTT based Forwarding (ASF) strategy was originally developed to aid Hyperbolic Routing which could create persistent suboptimal paths. The strategy adapts to network dynamics by utilizing NDN’s stateful forwarding and actively probing the data plane. However, ASF was designed for generally stable networks with static nodes. Therefore, it does not perform well in the presence of mobile nodes, where network topology varies frequently. In this work, we propose an improved ASF strategy for mobile networks, called m-ASF. The proposed strategy uses multiple paths when the primary path is experiencing failures. However, to prevent premature and frequent changing of faces, m-ASF employs anti-oscillation mechanism. We have done a comparative analysis of m-ASF and ASF. Our results show that m-ASF is more e than ASF in retrieving data in dynamic environment. The performance gain is attributed to m-ASF’s faster exploration of new paths and granularity in face-ranking. NDN-DPDK: NDN Forwarding at 100 Gbps on Commodity Hardware Junxiao Shi NIST Davide Pesavento NIST Benmohamed NIST Since the NDN data plane requires name-based lookup of potentially large tables using variable-length hierarchical names as well as per-packet state updates, achieving high-speed NDN forwarding remains a challenge. In order to address this gap, we developed a high-performance NDN router capable of reaching forwarding rates higher than 100 Gbps while running on commodity hardware. In this paper we present our design and discuss its We achieved this performance through several optimization techniques that include adopting better algorithms and t data structures, as well as making use of the parallelism by modern multi-core CPUs and multiple hardware queues with user-space drivers for kernel bypass. Our open-source forwarder is the software implementation of NDN to exceed 100 Gbps throughput, while supporting the full protocol semantics. We also present the results of extensive benchmarking carried out to assess a number of performance dimensions and to diagnose the current bottlenecks in the packet processing pipeline for future scalability enhancements. Finally, we identify future work which includes hardware-assisted ingress dispatching, dynamic load balancing across parallel forwarding threads, and novel caching solutions to accommodate on-disk content stores. Session 4: DARPA SHARE DARPA Secure Handhelds on Assured Resilient networks at the tactical Edge (SHARE) Mary Schurgot DARPA The DARPA Secure Handhelds on Assured Resilient networks at the tactical Edge (SHARE) program is developing security and networking architectures and software for sharing information across multiple security levels. SHARE addresses several technology challenges to enable tactical edge, small unit information sharing between coalition and U.S. forces. Sp, SHARE employs named data networking (NDN) to provide resilient secure store and forward capabilities to overcome brittle end-to-end connections. 5
PAGE – 6 ============
PLI-Sync: Prefetch Loss-Insensitive Sync for NDN Group Streaming Yi Hu Perspecta Labs Constantin Serban Perspecta Labs Lan Wang University of Memphis Alex Afanasiev FIU Lixia Zhang UCLA In this paper we explore solutions to robust group communication in disadvantaged wireless networks, which exhibit low bandwidths, high packet loss, and frequent/permanent network partitions. More specif -ically, we propose a group communication protocol based on Named Data Networking (NDN). By design NDN’s in-network caching and stateful forwarding plane can help improve data delivery robustness in disad-vantaged networks, however, when content is generated at a non-deterministic rate, an cient, low-overhead synchronization protocol is needed to inform group members of the new content to fetch. To address this need, we propose Prefetch Loss-Insensitive Sync (PLI-Sync) protocol, specialized for group communication in highly disadvantaged networks. PLI-Sync combines optimistic content prefetching with selective con-tent and addresses the challenges of distinguishing wireless packet losses from mobility-induced disconnections, and between data availability and retrievability. Our evaluations show that leveraging the interplay between NDN’s stateful data plane and a low rate sync protocol can sitly reduce commu -nication overhead compared to solely relying on sync protocols, while maintaining low data transfer latency and robust delivery in a variety of wireless conditions and load settings. NDN in DARPA SHARE John DeHart Washington University in St. Louis We give an update on our team’s progress in the DARPA SHARE project and highlight some of the mo and additions to NDN we have made. Some of the unique aspects of networking with tactical radios have lead us to some application and topology sp mo to overcome challenges in inter -mittent, lossy and at times segmented network environments. These mod include a new NDN sync protocol (ICT-Sync), and new NFD strategies to help reduce the network load. Session 5: IoT/Edge The \Decision Maker”: NDN concepts for intelligent automation Marie-Jose Montpetit Concordia University Montreal IoT systems are fast becoming data driven intelligent system that can accomplish real time tasks but also combine with edge and cloud computing to deliver complex decision-making. This is enabled by the availability of powerful sensors, cameras and edge devices that constantly communicate among themselves and with other systems in the \cloud-edge” continuum. The move to data networking is happening across a wide swath of potential applications from intelligent cars to self managed farming systems. Yet, not all data captured by the sensors is at the same level of importance or dedicated to the same decision system. Also, the networked micro-controllers are better addressed by name than by obscure addresses. The presentation will introduce the \Decision Maker” a layered NDN enabled architecture for distributed intelligent automation. \Decision maker” combines named entities, interest/data information and discovery as well as semantic web concepts to provide a framework for next generation IoT. The presentation will focus on the main elements of the architecture as well as introducing use cases to show its usefulness. 6
PAGE – 8 ============
(\generations”), maintaining coded caches, performing decoding and recoding in the routers, and establishing conventions for names. For high-level semantics, MICN revisits subtle questions such as merging similar Interests from dierent users: our current design chooses not to do so. An optional interest cancellation mechanism is also introduced. Then support for MILIC-sp semantics requires additions: adding an index in Interest names, gener-ating coded Data packets satisfying the MILIC constraints, pipelining multiple Interests, and proper support in case of packet losses. Finally, getting protocol details right requires MICN is no exception. We found that a critical part was introducing careful queue management for pending PIT entries. We will also discuss the remaining issues and on-going work, such as the integration of sophisticated interest forwarding strategies. References: [1] H Malik, C Adjih, C Weidmann, and M MICN: a Network Coding Protocol for ICN with Multiple Distinct Interests per Generation.arXiv preprint arXiv:2007.01128, 2020. [2] M.-J. Montpetit, C. Westphal, and D. Trossen. Network coding meets information-centric networking: An architectural case for information dispersion through native network coding.MobiHoc, (2):31{36, 2012. [3] J. Saltarin, E. Bourtsoulatze, N. Thomos, and T. Braun. Adaptive Video Streaming With Network Coding Enabled Named Data Networking. IEEE transactions on multimedia, 19(10):2182{2196, 2017. A Named Data Networking Architecture Design to Internet of Under-water Things Qi Zhao University of California, Los Angeles Zheng Peng The City College of New York Xiaoyan Hong The University of Alabama The Internet of Underwater Things (IoUT) advances our ability in exploring oceans, lakes and rivers through multiple communication technologies that connect stationary and mobile nodes underwater, at the surface and in the sky. However, characteristics such as low data rate, long propagation delay, energy-constraint, device mobility and sparsity, etc. of underwater communications remain as major challenges to the potential b that IoUT can bring to data availability and data sciences. This paper mainly focuses on further exploring how the Named Data Networking (NDN), a future Internet architecture, addresses the challenges of IoUT and can be adapted to potentially provide a cient, and secure implementa -tion of IUWT. The IoUT network and application semantics are aligned with the data-centric communication model of NDN, which enables operators to deploy and networks more easily, and developers to focus more on \things” and data underwater. The paper starts from introducing new challenges in IoUT and then illustrating in detail with simple examples to show how to employ NDN architecture to IoUT and how to enable additional functionalities required by IoUT. We also elaborate on the detail of our way of thinking and the remaining challenges while applying NDN to IoUT scenarios for future research directions. Improving Existing Software Applications with a Practical and Secure NDN Publish/Subscribe Transport Randy King Operant Networks Thompson Operant Networks Kathleen Nichols Pollere Inc. Operant Networks is a startup founded in 2015 to initially address the problem of unreliable and insecure communications to renewable energy sites through NDN. We have since broadened our scope to include a wide range of energy industry applications, all with a strong focus on cybersecurity. Working in partnership with Pollere, UCLA, US Air Force, US Department of Energy, and key industry partners we have developed 8
PAGE – 9 ============
an interesting NDN-based transport to replace an existing TCP/IP publication/subscribe framework in a cybersecurity intrusion detection system. Historically, industrial networks are secured through perimeter alls to exclude malicious It is well known that these are fraught with y and unlikely to be completely e at excluding determined attacks. In response, intrusion detection systems (`IDS’) have become a growing market segment, aided by two realizations: Nearly all network is encrypted and to inspect in a all Advanced hackers tend to lurk in networks for weeks, learning the details of the site, before imple -menting their attack; allowing time to detect their presence Probably the most advanced IDS systems are based on distributed internal sensors: network packet deployed throughout a campus to monitor internal as well as that leaving the site. A leading example of this type of IDS is Zeek (zeek.org), developed over decades as an open source project. It was developed from early on as a distributed system; a cluster of multiple packet instances on multiple physical servers collaborate to aggregate and analyze large amounts of distributed information. Recently, Zeek re-architected their IDS framework to utilize a publish/subscribe model over discrete TCP/IP links secured by SSH. This communications paradigm provided Operant an opportunity to replace it with an NDN-based multicast pub/sub transport. In addition to the known security b of NDN, this transport demonstrates the commercial use of grained and deployable trust schemas within a pub/sub framework. Our belief is that the intrusion monitoring solution must be much more secure than the underlying customer network or it provides little added security value. Additionally, the NDN transport utilizes VLAN UDP multicast to remove many of the practical headaches of installing a distributed system within a TCP/IP network; reducing the need to maintain certis or static IP addresses. A large IDS installation might utilize hundreds of sensor appliances, this could become an overwhelming maintenance burden in the TCP/IP world. Most generally, we have learned that deploying NDN in an existing communications model (pub-lish/subscribe) is a much easier path to commercialization than switching applications or their users from a TCP-based model to an Interest/Data exchange paradigm. We implemented the NDN Pub/Sub with an MQTT-like API (the most popular IoT Pub/Sub) and look forward to discovering other existing applications that can b from our lightweight secure transport and its complex trust models and broker-less pub-sub framework. Session 6: Security/MAC Rolling out NDN for DDoS Mitigation Zhiyi Zhang UCLA Sichen Song UCLA Angelos Stavrou George Mason University Eric Osterweil George Mason University Lixia Zhang UCLA Distributed Denial of Service (DDoS) attacks have plagued the Internet for decades, but the basic defense approaches have not fundamentally changed. Rather, the size and rate of growth in attacks have actually outpaced carriers’ and DDoS mitigation services’ growth, calling for new solutions that can be, partially or fully, deployed imminently and exhibit eness. In this talk, we examine the basic functions in Named Data Networking (NDN), a newly proposed Internet architecture, that can address the principle weaknesses in today’s IP networks. We demonstrate by a new DDoS mitigation solution over NDN, Fine-grained Interest T Throttling FITT, that NDN’s architectural changes, even when incrementally deployed, can make DDoS attacks funda-mentally more to launch and less e. FITT leverages the NDN design to enable the network to 9
PAGE – 10 ============
detect DDoS from victim’s feedback, throttles DDoS by reverse its exact paths through the network, and enforces control over the misbehaving entities at their sources. Our extensive simulation results show that FITT can throttle attack with one-way time delay from the victim to the NDN gateway; upon activation, FITT ely stop attack from impacting benign ws, resulting in over 99% of packets reaching victims being legitimate ones. We further demonstrate that service providers may implement NDN/FITT on existing CDN nodes as an incrementally deployable solution to the application-level remediation at the sources, which remains unattainable in today’s DDoS mitigation approaches. Secure Sharing of Spatio-temporal Data through Name-based Access Control Laqin Fan University of Memphis Lan Wang University of Memphis Named Data Networking (NDN) is proposed as a future Internet architecture, which provides name-based data publishing and fetching primitive. Compared to TCP/IP, NDN removes the need to manage IP address and has semantically meaningful names, NDN has stateful and name-based forwarding which the network stack, NDN’s data-centric security and in-network caching are more t than cloud-based data delivery and encrypted channel. Name-based Access Control (NAC) is a content-based access control in NDN, which requires access control by encrypting content at the time of production directly without relying on a third-party (i.e., Cloud) to host the contents, and utilizes NDN’s hierarchical naming convention to express access control policy and accomplish automating key distribution. As more sensitive data is generated and stored, there will be a need to share the data securely, such as mobile health (mHealth) data, data from smart-home systems. These data are generated over time and/or location, and could be collected as ongoing data streams. The data owners may want to share entire data or a time-location-sp subset of the data based on their preferences. Therefore, access control to those data must address the spatio-temporal attributes, and support secure real-time data sharing as well. An e and secure access control solution is required to ensure that only authorized users can access to certain data. Inspired by Named-based Access Control model with data-centric security, we take into account the data attributes to make access decisions. By specifying access control policies with time and/or location attributes, we could limit data access to a given time bound and/or location area. In this work, we make three contributions: (1) we design an NDN’s semantically structured naming convention to express access control policy on spatio-temporal data, (2) for real-time data sharing, we deploy PSync to have a publish-subscribe module, (3) we have implemented a practical spatial-temporal data access control prototype based on NAC library in NDN codebase. Moreover, we evaluate the performance using Mini-NDN for t content key granularity. A Full Data-centric Network Stack Integrating V-MAC and NFD Mohammed Elbadry Stony Brook University Fan Ye Stony Brook University Peter Milder Stony Brook University YuanYuan Yang Stony Brook University Current NFD performance on wireless networks is severely obstructed by the lack of a true data-centric MAC layer. Building on top of V-MAC, our novel data-centric MAC prototype that ers high rate (up to 65Mbps on pi and 900 mbps on another platform), low loss (1-3%) multicast support (dozens of receivers), we present V-MAC NFD Link Protocol (V-NLP), a Link Protocol integrating V-MAC and NFD for a fully data-centric network stack. The new stack eliminates the network grouping concept, and enables pub/sub abstraction at the radio level. V-NLP allows researchers and developers to run NFD over V-MAC with multiple frame types (Interest, Data, Announcement, Implicit Interest), customized data rates, and multicast support. This fully data-centric stack will enable NFD to achieve high performance 10
PAGE – 11 ============
for applications in vehicles, drones and IoT environments. The source code for V-NLP, V-MAC and ware for supported chipsets will be released to the community. Session 7: Plug-n-Play NDN Eric Newberry UCLA Tianyuan Yu UCLA Zhiyi Zhang UCLA John Dellaverson UCLA Lixia Zhang UCLA A major barrier to the adoption of NDN is the high manual required to set up and a functioning NDN network consisting of multiple hosts. To date, such has required manual on every host to establish initial content reachability and manual of links between forwarders whenever there is a topology change. These cguration tasks are particularly when one is working with embedded or IoT devices. Meanwhile, with IP networks, one can connect multiple end devices together in a local network and have them \just work”. Therefore, to help encourage the adoption of NDN beyond the research community, we have developed \Plug-n-Play NDN”, where an NDN environment can easily established in a LAN environment. We demonstrate its eness using a small example scenario. NDNSD: Service Discovery in NDN Saurab Dulal The University of Memphis Lan Wang The University of Memphis Service discovery (SD) is a fundamental requirement of contemporary networking systems. modern web, IoT application, building management, smart device, LAN, WAN, mobile application, etc depends on SD in a way or another. The proliferation of the applications towards the edge, rapid expansion and advancement of IoT and sensor networks, cloud computing, etc have pushed the service discovery to the next level of importance. It is also changing the dynamics of how the services and resources were discovered and used in the past. Furthermore, the pervasive nature of wireless networks and mobile devices has underscored its importance and is expected to be even more in the future. This presentation introduces NDNSD, a general-purpose, fully distributed, and scalable service publishing and discovery mechanism for NDN using the NDN synchronization protocol. Data synchronization protocol plays a crucial role in NDN. The original NDN architecture envisioned combining sync protocol with application accessible libraries to provide transport functionalities to the applications. The application accessible libraries should hide the core network functionalities and primitives such as interest and data from the applications, and the sync should help the transfer of data from one application to another. The Internet protocol stack is a well-known example of such a model { best known as the hourglass model. Remaining in the realm of the internet hourglass model, releasing the importance of SD in NDN, and taking inspiration from some of the previous works, we design and developed NDNSD. It is an application accessible reusable library that uses synchronization protocol for service announcements and discovery. We view service discovery problems as data synchronization problems. Some applications actively looking to discover services while others trying to advertise the services. There are several b of using sync for SD i) no external dependencies or demanding change in network layer { sync comes with NDN natively ii) inherently supports multi-way communications, iii) Flexibility to implement application semantic and so on. Furthermore, two or more parties can agree on a common sync group. This applies to local as well as global applications or devices. A mobile application can agree on sync group \/letschat”, whereas printer services can agree on \/printers”. Similarly, IoT and edge applications can have their sync 11
143 KB – 16 Pages