Merging engine implementation for intra-frame sharing in multi-tenant virtual passive optical networks

Akhlaque Ahmad, Ashfaq Ahmed, Arafat Al-Dweik, Syed Taha Ali, Arsalan Ahmad

    Research output: Contribution to journalArticlepeer-review

    2 Scopus citations

    Abstract

    High up-front capital expenditures impede the widespread deployment of passive optical networks (PONs). A multi-tenant solution, in which multiple network operators virtually share PON infrastructure and bandwidth resources, can result in significant cost savings. First, we investigate the viability of virtual PONs, in which virtual network operators share a portion of the upstream bandwidth in a frame. Each PON schedules its frame-level capacity using a dedicated dynamic bandwidth assignment algorithm, resulting in the generation of an independent bandwidth map (BMap). Second, we implement at the optical line terminal a merging engine that combines individual virtual BMaps into a single physical BMap, which is then transmitted to all optical network units. Our novel traffic merging algorithm is capable of consolidating traffic across all transmission containers. We apply this merging engine on top of the 10-Gbit-capable PON module and validate our approach using network simulator 3. The results show that our proposed intra-frame sharing technique, when integrated with our traffic merging algorithm, significantly outperforms the standard inter-frame sharing approach in terms of both latency and packet loss. Furthermore, we observed that best-effort traffic is affected most for all resource constraint scenarios of intra-frame sharing.

    Original languageBritish English
    Pages (from-to)209-218
    Number of pages10
    JournalJournal of Optical Communications and Networking
    Volume15
    Issue number4
    DOIs
    StatePublished - 1 Apr 2023

    Fingerprint

    Dive into the research topics of 'Merging engine implementation for intra-frame sharing in multi-tenant virtual passive optical networks'. Together they form a unique fingerprint.

    Cite this