Multidomain SDN-Based Gateways and Border Gateway Protocol

Table 3

Comparison of published works.

Proposed framework	Description	Remarks/Limitations	Base protocol	Difference from other published works
ETMP-BGP [15]	A framework that integrates end-to-end AS-level paths in domains	Focused on enhancing the connectivity in terms of the iBGP not eBGP	BGP-4	It focuses on enhancing the connectivity in terms of eBGP not iBGP.
INDOPRONOS [27]	Method and techniques to support automatic provisioning in multidomain SDN via east/west interfaces	It has enhanced the connectivity between multidomain SDN in terms of reachability, but it has a high BGP-4 convergence delay. Also, it does not fully evaluate scalability and resiliency.	BGP-4	It focuses on how to enhance and minimize the high BGP-4 convergence delay and the scalability scenarios.
SDNi [54]	An IETF draft standard that proposed an automated system protocol to connect multidomain SDN	It does not have any work progress since December 2012.	BGP-4 or SIP	It focuses on multidomain SDN technology using the BGP-4 routing protocol updated.
FASTPLANE [77]	A fast BGP-4 simulation tool for large data centres	It is effective in small to medium size data centres, but not for large data centres. It has some validation of BGP-4 semantics and production bugs in routing policy, network architecture, and router firmware.	BGP-4	It focuses on minimizing the BGP-4 convergence delay on large data centres.
DISCO [79]	It is a distributed SDN control plane for communication between multidomain based on AMQP protocol.	It uses a Floodlight controller and mainly focuses on connectivity and reachability, not QoS.	AMQP, but use BGP-4 Agent for Interoperability	It focuses on connectivity and reachability, keeping the QoS in the packet loss without measuring the BGP convergence delay and scalability parameters.
EMPLaaS [14]	framework to evaluate the MPLS in an SDN environment in the control plane layer.	It been evaluated its performance in a virtual environment not real-world deployment.	BGP-4	It evaluates in a real network environment at RMIT University lab.
SDX [16]	Interconnect SDN with legacy IP Network	SDN-based Internet exchange. It does not test the approach in multidomain SDN.	BGP-4	It focuses on multidomain SDN networks
(Bassey and Nayak, 2018) [13]	They proposed a framework for iBGP environment.	The proposed framework increased the BGP-4 convergence time 30% and 17% compared with the baseline results.	BGP-4	It focuses on minimizing the BGP-4 convergence time.
ONOS SDN-IP [82]	BGP-4 interfacing applications, which enables multidomain SDN and can interconnect SDN domain to legacy IP network.	Between SDN-IP applications or between SDN-IP and legacy IP network not between SDN-IP and SDN-IP networks.	BGP-4	It focuses on multidomain SDN networks.
(Zhou, Wu, Cheng, & Liu, 2017) [83]	It observed to investigating the BGP-4 convergence issue by means of interdomain multipath routing that distributes the routes and tackles the overhead messages.	The model needs more effective since they focused to get improvement in service availability and packet loss without any BGP-4 convergence delay improvements.	BGP-4	It focuses to get improvement in network scalability and BGP-4 convergence delay.
An Avoid Loop with Test Packet (ALTP) scheme [84]	A framework with multicontroller architecture that dodges packet drops and forwarding loops without multiple-domain controllers’ synchronization in the network using ALTP scheme.	It focuses on reachability issues and it showed an improvement in minimizing the latency of the flow setup and utilizing the workload of the controller comparing with the baseline framework, furthermore, link-load balancing facilitates the improvement of the throughput, under multiple domains. However, it has a limitation in recovery of control layer failure which makes the entire domain disconnected from the other domain.	OSPF, probe test packet scheme, and Link Layer Discovery Protocol	This review is different because it uses BGP-4 as a routing protocol between different SDN domains in different AS, and it focuses on minimizing the BGP-4 convergence delay.