Ring Topologies - Fault Protection
Updated: Dec 16, 2020
Hussle between Ethernet Topologies
The Fast-growing demand for high-speed internet is fulfilled by building an Ethernet Network. Historically, speaking the word "ether" of ethernet is a way of describing an essential feature of the system; the physical medium carrying bits to stations as described by its inventor, Metcalfe. Ethernet, which was initially launched as a straight line or bus network topology to provide an internet connection to home users only, now provides a connection to multiple enterprises, home, etc. networks in different topologies such as a star, bus, ring, etc. The Selection of right topology ensures the high performance of a network, easy manageability of network operations, and increased data efficiency with supporting the economics of it. In all the topologies, the most common and prominent structure is the Ring. The Ethernet Ring Topology also provides a redundant infrastructure that is necessary in today's world.
What exactly is a Ring Topology?
A network configuration in which devices in a circular path are connected to facilitate data flow is a Ring Topology. Each network device connects with two others, like points on a circle, these network devices combined, make a ring network. The data packets in a ring network travel from one device to the next until they reach their destination. The Ring topologies provide budget based expected end-user performance in a complex IT Landscape. Now, let's see how Ethernet embarked its journey with Ring Topologies
Given the commercialization of Ethernet in the 1980s, it has come a long way. In the past few years, the demand for ethernet has increased dramatically. To support the high traffic demand in different areas, the networks have gotten more and more complex and to ensure the efficient operation of such complex networks, it is necessary to have a fault-proof network. The Network Topology of Ethernet Rings provides an economical and safe solution of a fault-proof network as compared to other ethernet topologies. Ethernet Ring systems are commonly known as Ethernet Ring Protection Switching. The Ethernet Ring Protection Switching system provides less convergence time regardless of traffic and size of the network as compared to other switching systems in case of ring faults.
As a part of a Ring Network, Ethernet is supposed to take care of the problem of looping. Data packet could circle a ring endlessly and consume precious bandwidth. Whenever there is more than one Layer 2 (OSI-Model) path between the two endpoints, a switching loop or bridge loop takes place in computer networks. The main issue that crops up due to a loop is the problem of rebroadcast of the broadcast messages, which leads to flooding in networks. This happens due to the incompetency of the loop that creates broadcast storms resulting in forwarding broadcasts and multicasts packets by switches of every port. As an example- 4 Devices - A, B, C, D are connected in a Ring, First packet starts from A with data to deliver it to D, the packet delivers the data but is not dropped (i.e., no loop prevention), it keeps flowing in the Ring. Similarly for the next packets such as 2nd, 3rd …etc, will also keep flowing in the ring without any purpose, this will create flooding of packets in the network.
Ethernet Protocols for Ring Topologies
Ideally, a switched network should not contain loops, yet due to the efficiency, economical & redundancy reasons. So how do we create a loop-free logical topology of Ethernet? The answer which first comes to our mind is using the shortest path bridging (SPB) protocol or using the older spanning tree protocols (STP) on the network switches. There are a lot of protocols that are in use, let's discuss in brief for the three most common in Ethernet networks - ERPS, MSTP (802.1s) and RSTP (802.1w). Here is a quick comparison of the pros and cons of the three most common protocols on the basis of their convergence time, complexity, etc. factors
RSTP has been the go-to method for building mesh networks due to its advantages over the older methods. It works by redundantly linking the ethernet networks, but its most significant limitation is that it takes between 2 to 10 seconds for detecting and responding to a failed link. The inability of not being able to handle more than 7 hops makes it an expensive choice for networks. MSTP, on the other hand, surpasses the RSTP for its ability of faster convergence times, which stems from its ability to map multiple VLANs on a single set of equipment, yet its 10- second convergence times is a bit too much. The convergence time of RSTP and MSTP is more because it needs extensive information exchange for tree computing. A basic STP Topology diagram is shown below.
To deal with the drawbacks of the above two protocols, ERPS was introduced. It has a clear advantage because of its low convergence time, and its universal nature. ERPS fits the existing network organically without any extra hardware costs It has a clear advantage with low convergence times, and with its universal nature, it fits the existing network organically without any extra hardware costs. The faster redundancy is an added benefit that provides more substantial scalability as there is no theoretical limit on the number of rings that can be added to a network.
G.8032 Ethernet Ring Protection Switching
Ethernet Ring Protection Switching (ERPS) is the most utilized approach in the network infrastructure by top tier companies such as Cisco, Juniper, Huawei, etc. Some of the most common industries that utilize G.8032 ethernet topology structure are mining, waste management, gas, power and utility monitoring, automated production facilities, surveillance and security, safety monitoring along with harsh-environment workplaces. ERPS enables a large amount of Ethernet traffic to flow to multiple connection pints with high-level redundancy. ERPS is an obvious choice for customers because of its beneficial aspect in which the ITU-T G.8032 management protects Ethernet traffic and maintains recovery times under 50 milliseconds (ms). The G.8032 prevents looping by actively managing traffic. When everything is working nominally, the protocol blocks traffic on a single ring link to prevent looping. If a node fails, the protocol reroutes traffic (opening the closed link as needed) to keep data flowing. When the failed node is restored, the protocol automatically closes the temporary link to continue protecting against data loops. The temporary link is known as Ring Protection Link (RPL). Let's look at it more closely.
Working of ERPS
For an ethernet ring operating in a healthy condition, the ethernet ring adjacent to RPL configures as RPL owner node, and another ethernet ring adjacent to RPL configures as RPL neighbor node. Ring Node D and Ring Node A configures as an owner node and neighbor node, as shown in the below Figure. Both the end nodes block the reception and transmission of data through RPL under normal operating conditions. When a failure occurs in Ring Node B and Ring Node C, the link between them is blocked, and the RPL Link starts working. See the below diagram and flow chart for reference.
Flow Chart for ERPS Mechanism
G.8032 Version 2 ERPS
Similarly, to the G.8032 Version 1, Version 2 operates on multiple rings, providing support for sub-ring topologies. Sub rings are "C" shaped rings, which are not fully closed. The sub-rings attach with nodes of a major ring. Some of the multiple ring networks are given below. Fault protection is similar to a single ring topology.
Intellectual Property analysis of Ethernet Rings
Let us look at how technical leaders have shaped the Intellectual Property space with ERPS Patents. We have broadly analyzed the List of granted patent families with top assignees in the ERPS domain. There are a total of 866 granted patents in the ERPS domain until today. Fujitsu leads with a close difference from other industry leaders. Here is the quick view of the analysis.
Furthermore, Lumenci has analyzed the different sub-topics in the Ethernet domain. As an example, there are 13 patents in the Loop-Free domain for Ethernet Ring Topologies.
The Road Ahead
Customers or users around the globe are always in a quest to seek a service that is seamless and saves them that little second or that millisecond of time, especially when it comes to internet access. ERPS, with its low convergence times based on the rapid restoration time of under 50ms, has become the obvious choice for the top tier network providing companies. The G.8032 ERPS is supported over the server and physical layers, supporting core metro and access network applications.
Building on the same, Lumenci has gone a step further to conduct an IP analysis that indicates trends in the ERPS for different subclasses & different vendors. To gain additional insights in the Ethernet Ring Protection, one can always refer to the vendor-specific ring solutions like Extreme Network’s Ethernet Automatic Protection Switching (EAPS), Cisco’s Resilient Ethernet Protocol (REP), Broadcom’s Metro Ring Protocol (MRP), to name a few. As it is said, “to each his own”, although there are various topologies, different protocols can be best suited as per the topology.
Chirayu Parashar Associate Consultant- Lumenci