The ring and the star: who is whom?

When building networks, there are two competing topologies: this is a star (of different variants) and a ring (of different variants). The star has one advantage - low re-subscription. The disadvantages of the star - a complex structure, and therefore the complexity of operation and high cost. A star is a solution for accessing users in one place: a classical network of the management building in an enterprise, but not always and even in this case. After all, the core network of the building in compliance with fire safety standards will always be laid through two cable risers in different parts of the building and this again is a ring, not a star: one ring / cable of optics through these two risers and two centers of the core of the building / data center network.

The choice of a particular solution topology depends on the object and its features. The ring is always more profitable for distributed networks, since it is very, very expensive and really impossible to make a star on a large distributed network. Therefore, ring topology is the optimal topology for large enterprises, processing plants, urban networks, networks of national scale.

Fig. 1. Module of the automated process control system of the plant.

All network equipment manufacturers can work on the star topology, and only a limited number of manufacturers — HP, Huawei, Extreme — can work on the ring topology on switches with the distribution of virtual networks throughout the campus without using complex and expensive MPLS / VPLS technologies.
')
For example, take a couple of typical simple "ring" objects: a stadium and an airport.

Fig. 2. Stadium network.

For these objects, the ring topology has the following advantages compared with a star:

1. It is much easier to install and maintain the ring topology. It is much easier to install and maintain the ring topology, since access devices through the ring immediately fall either into the server farm or into the core, and the star still assumes an intermediate level - aggregation of channels from the access level. The reason for the aggregation level of a star is simple: at the aggregation level, the cost per port is much lower than at the star level of the star, as well as for more flexible application of various policies.
2. For modern networks, the reservation of communication channels from the access level to the aggregation level and / or core is critical. If a star makes reservations on cable routes along different paths, then physically it will turn out to be a ring, and logically it will be a star. But at the same time, for each switch on access, it will be necessary to drag individual cable routes to the switching points, since the optics are always cut and welded with the whole bundle, and not along individual conductors. Each welding optics on the path of light in the fiber increases the loss of the budget of the optics, and as a result, reduces the distance of the optical communication channels. It should also be borne in mind that the work on welding optics will be several times more with the star topology than with the ring topology.
3. For stadiums, airports and networks of enterprises there are always at least two separate physically separated networks. For a ring it is simple, for a star with distributed objects - the reality is sad.
4. The convergence of the star will always be worse, since there will always be loops of logical paths between the access switches, except in the case when in the core two slot switches work as one. The convergence of the ring topology is from 50 ms (one ring) to 200 ms (subrings are connected to the main ring).
5. At the star of difficulty with scaling: the addition of cable lines in any place is a broach of a new additional cable, for a ring there is an additional coupling in the existing optical bundle.
6. In the case of a ring, we always remove the necessary and planned uplink speeds from the switches that are most loaded from access to the core: that is, we pay for the increase in bandwidth as needed. And in the case of a star, we pay for everything at once and really do not use the band of connected uplink channels.

The last point I would like to consider in more detail. It should be noted that the connection of servers will always be a bottleneck: since if we connect 28 (this is more than enough for even the world's largest stadiums) gigabit switches in star topology, then we have an overload when connecting access to the core will be 2 * 2 * 20GB (speed of connection of the server module to rings) / 28 * 20GB (28 access switches, two for each 10 GB / s uplink) = 1: 7. In the case of a stack, this will be 40GB (two server stacks) / 4 * 20GB (four access stacks) = 1: 2. But a re-subscription on access will have a star of 20GB (two uplinks of 10GB each) / 48 * 1GB (access ports) = 5:12 = 1: 2.4, and a stack of 20GB (stack width) / 7 * 48 * 1GB (seven access switches in one stack) = 5: 7 * 12 = 5:84 = 1:17. For a standard network access, overload is allowed 1:20. As we can see, in the case of a star, a low access overload causes a “choking” for servers. To solve these problems in the network core for the star topology, you need to install expensive modules (but the question is, why add them, if server ports are already enough?), If you need to reduce congestion on the access in the stack, you just need to add optical converters. Below is a comparison table.

Technology	Stek	Star	Remarks
Re-subscription on access	1:17	1: 2	Valid 1:20
Reduce re-subscription on access	A pair of optical converters	-	The cost of each SFP + low - 3K. But, given the experience of building the previous stadiums, where access is 100 MB in general, this overload will be more than enough.
Reducing oversubscription when connecting servers	-	Server modules in a modular switch	Question: if there are only up to 50 servers, why connect more ports, half of which will not be involved even theoretically ?!

But here the progress interfered with the ring and the stars: 40 GB and 100 GB interfaces appeared. For 95% of clients, these are simply unreachable in the near future uplink communication channels for connecting to the kernel. And as a result, for the same veins, without adding anything, but simply replacing the 10 GB transceiver with 40GB or 100GB, we increase the bandwidth of the ring several times, and then the potential problem of oversubscription of the ring does not appear in principle. Although, again, the price of 40 GB and 100 GB transceivers per 10 km is very high now, but in a year it will be the price of current 10 GB transceivers.

The largest examples of the use of ring technologies in conjunction with the stacking technology in Asia are the city of Beijing, in Europe the French railways. But the French railways simply had no choice after the TR "died." And these are projects implemented on HP equipment (former 3Com equipment), which gives reason for reflection and practical application of the described ring design for data network designers.