Telecommun Syst (2012) 49:299–312 DOI 10.1007/s11235-010-9376-1 Scalable, hierarchical, Ethernet transport network architecture (HETNA) Eliav Menachi · Chen Avin · Ran Giladi Published online: 29 June 2010 © Springer Science+Business Media, LLC 2010 Abstract Ethernet technology is not inherently scalable, and yet, Ethernet dominates LANs, and recently has dif- fused to access, aggregation networks, and MANs. Ether- net is even considered for transport networks in the back- bone. Many solutions from IEEE, IETF, and MEF are con- sidered for enabling Ethernet beyond LANs and bridged LANs. In this paper we offer HETNA, a hierarchical Eth- ernet forwarding, which is compatible with legacy networks used by enterprises, carriers, and backbone networks, and provides scalability, mobility, protection, multicasting, and QoS support to intra- and inter-domain networks in an ef- ficient forwarding manner. The suggested architecture can handle streaming, real-time, multicasting, and other appli- cations as well as various addressing mechanisms (e.g., IP or URI addressing). Both connection-oriented transport ser- vices and connectionless-oriented services are possible in the suggested architecture. This architecture was simulated and prototyped, showing significant improvements over reg- ular Ethernet in terms of buffers and control messages that enable this network to function. Keywords Ethernet · Scalable · Mobile · Inter-domain · Intra-domain This research was funded by the European Community Seventh Framework Programme [FP7/2007-2013] under grant agreement 215462. E. Menachi · C. Avin · R. Giladi ( ) Ben-Gurion University of the Negev, Beer-Sheva, Israel e-mail: ran@bgu.ac.il E. Menachi e-mail: menachi@bgu.ac.il C. Avin e-mail: avin@bgu.ac.il 1 Introduction During its four decades, Ethernet technology has become the dominant LAN technology, and its bandwidth capacity increases about twice every two years, on average. During this period, Ethernet adopted many modifications that han- dle its speed, media, bridging functionalities, segmentation and virtual LANs, and more. Currently, many attempts are being made to expand this inexpensive and efficient tech- nology beyond home and enterprise networks into the ac- cess, aggregation, metro, and even core networks. However, maintaining the basic Ethernet concept, addressing and for- warding mechanisms restrict Ethernet from being fully scal- able, and confine it mainly to home and enterprise networks. This is due to the use of “flat” Ethernet MAC addresses, which leads to the need to store and maintain all MAC ad- dresses in every forwarding node, and broadcasting frames with unknown addresses by the forwarding elements. Eth- ernet also lacks some basic networking capabilities that are required in contemporary networks, certainly in the aggrega- tion, metro, and transport networks, e.g., mobility and QoS support. There are currently many attempts to address the above Ethernet drawbacks by standardization bodies (i.e., IEEE, IETF, MEF) and researchers; however, the scalability issue remains the major obstacle in using Ethernet ubiquitously for all networking purposes [4, 7, 13, 17]. Although bridges and virtual LANs (VLANs) are supposed to handle scalabil- ity, they do not really solve the issue. Bridges were orig- inally used to separate LAN segments, to facilitate inter- connection of various LAN technologies, to include more connected hosts, and to better utilize isolated segments [6]. Bridges either forward frames destined to specific hosts ac- cording to the hosts’ addresses and locations, or they broad- cast frames with unknown destination addresses. A bridge