This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| ARP-Path | |
|---|---|
| Name | ARP-Path |
| Developer | University of Lisbon |
| Introduced | 2010s |
| Layer | Data Link Layer |
| Primary use | Layer 2 forwarding, loop avoidance |
ARP-Path
ARP-Path is a layer‑2 forwarding technique designed to build low-latency, loop‑free paths by exploiting the Address Resolution Protocol and flooding dynamics across Ethernet switches. The approach was proposed to combine ideas from spanning tree algorithms and dynamic path learning similar to techniques used in mesh and metro networks, aiming to improve throughput and resilience in campus and data center environments. The design draws inspiration from research groups and projects associated with institutions such as University of Lisbon, Universidade de Coimbra, Carnegie Mellon University, Cisco Systems, and standards efforts like IEEE 802.1D, while engaging with use cases from deployments by organizations like Internet2, GEANT, and Amazon Web Services.
ARP-Path was introduced as an alternative to spanning tree protocols exemplified by IEEE 802.1D and innovations like TRILL and Shortest Path Bridging (SPB) developed by IETF and IEEE. The technique leverages Address Resolution Protocol behavior standardized in RFC 826 and operational models studied by teams at MIT, Stanford University, and University of California, Berkeley to dynamically establish active forwarding entries. Early publications and demonstrations involved collaborations among researchers affiliated with ACM SIGCOMM, IEEE INFOCOM, and conferences such as USENIX, HotNets, and IFIP Networking.
Motivations for ARP-Path trace to limitations observed in protocols like Spanning Tree Protocol (STP), Rapid Spanning Tree Protocol (RSTP), and proposals from IETF for layer‑2 scalability, as well as concerns raised by operators at Google, Facebook, and Microsoft about fault recovery and latency. Research compared ARP-Path goals with approaches from Shortest Path Bridging (SPB) work by IEEE 802.1aq and TRILL designed at IETF TRILL WG, reflecting influences from academic studies at Princeton University, ETH Zurich, and École Polytechnique Fédérale de Lausanne. The context included practical needs demonstrated in testbeds like PlanetLab, Grid5000, and production clouds operated by Rackspace.
ARP-Path operates by having switches learn best paths through observation of ARP request and reply packets following behavior specified in RFC 826, with mechanisms conceptually related to techniques used by OpenFlow controllers and routing protocols such as OSPF and IS-IS for path selection. Packets are flooded and the first arriving ARP reply creates forwarding state similar in effect to shortest‑path installation strategies studied in Bell Labs research and projects at Bellcore. The protocol avoids loops without disabling multiple physical links by using timed learning and aging rules analogous to those in Link Aggregation Control Protocol (LACP) used by IEEE 802.1AX, while incorporating failover responses comparable to fast reroute mechanisms from IETF drafts and operator practices at Verizon, AT&T, and China Telecom.
Implementations were prototyped on commodity switches and software platforms such as Open vSwitch, Linux kernel, and vendor gear from Broadcom, Intel, and Arista Networks. Experimental deployments used testbeds including Emulab, Mininet, and collaboration environments at CERN and National Institute of Standards and Technology. Integration efforts referenced management frameworks like SNMP and orchestration systems developed by Red Hat, Canonical, and VMware. Vendors and researchers compared ARP-Path implementations with proprietary features available from Juniper Networks and HP Enterprise.
Evaluations measured latency, throughput, and convergence time using benchmarks and methodologies from SPEC, NetPerf, and academic studies from Columbia University and Cornell University. Results showed improvements in path utilization and failover speed relative to legacy STP in scenarios modeled after topologies used in Google and Facebook data centers, while exhibiting tradeoffs similar to those reported for TRILL and SPB in publications at ACM SIGCOMM and IEEE INFOCOM. Simulations used tools and datasets from ns‑3, OMNeT++, and network traces collected by CAIDA and RIPE NCC.
Security analyses considered threats familiar to operators of Cisco Systems and Juniper Networks equipment, including ARP spoofing and poisoning described in advisories from CERT and vulnerability reports by SecurityFocus and MITRE (CVE database). Mitigations draw on countermeasures such as static ARP entries, port security features used in HPE switches, DHCP snooping and dynamic ARP inspection implemented in Cisco IOS, and cryptographic proposals discussed at IETF working groups. Threat models referenced incidents involving organizations like Target Corporation and Equifax to illustrate risks from layer‑2 exploitation.
Comparative analyses positioned ARP-Path alongside alternatives including Spanning Tree Protocol (STP), Rapid Spanning Tree Protocol (RSTP), TRILL, Shortest Path Bridging (SPB), and controller‑based solutions built with OpenFlow and SDN controllers from ONOS and OpenDaylight. Trade‑offs mirror discussions in standards forums such as IEEE 802.1 and IETF about complexity, interoperability, and hardware support, referenced in operational reports from Amazon Web Services, Microsoft Azure, and research projects at Yahoo! Research and IBM Research.
Category:Network protocols