Redundancy in a network is important to increase reliability. However, it does come with disadvantages such as broadcast storms. These disadvantages are offset by this Spanning Tree Protocol (STP) because it prevents so-called loops that arise due to the required redundancy in a network.
In this article we explain what broadcast storms are and how exactly the spanning tree protocol prevents them.
The term "broadcast storm" describes a problem in networks where a large number of broadcast messages overwhelms network traffic and overwhelms network resources. The most common cause of this is redundant cabling with two or more uplinks between two switches. In such a case, broadcasts and multicasts are forwarded to all ports except the port from which the traffic came. This creates a loop (switch loop) and the switches forward the broadcasts of the respective other switch.
Broadcast frames are output by the switch on all ports. The frames are duplicated and created via redundant connections Endless loops (endless loop/ network loop/ switching loop), which load the network so heavily that normal operation no longer works.
If network redundancies cause problems, why not just avoid them?
The benefit of redundant connections is the Avoiding a single point of failure: Redundant connections are set up in local networks to increase fault tolerance in the network and create backup options. In this way, the reliability of a network can be guaranteed.
In order to be able to use the advantages of redundancies in networks without accepting the disadvantages, the spanning tree protocol is used. In the following sections, we'll take a closer look at how this works.
The Spanning Tree Protocol, or STP for short, was standardized in 1990 in the IEEE standard 802.1.D. It works on layer 2 of the OSI layer model (data link layer) and prevents so-called loops that occur due to the required redundancy in a network. The STP also prevents other problems caused by redundancies, such as address table inconsistencies and frame duplication. To achieve this, the STP uses so-called Bridge Priority Data Units (BPDU) for the exchange of information in a network.
At BPDU These are packets that switches exchange with each other in order to use the transmitted information to establish a reliable network topology and hierarchy. The transmission paths in local networks are kept clear despite redundancy and meshed structure. A BPDU contains information about switch ports such as port ID, port priority and MAC addresses, which are relevant later for determining the root bridge.
BPDU TCN Topology Change Notification
BPDU TC topology change
BPDU Conf/Hello
Before we get into the function of the STP, it is worth mentioning its versions that have been developed over the years.
This is a list of all versions of the spanning tree protocol:
A so-called root bridge can be selected using the STP algorithm. As already mentioned, the spanning tree protocol dates back to the early 90s. At that time, switches were not that common. Instead, they worked with so-called bridges. The root bridge is selected via the so-called bridge ID, which contains the following information:
The switch with the highest/best priority (lowest value) is made the root bridge. However, it is quite possible that several switches have the same priority. According to the spanning tree protocol, the following then applies: Whoever can show the "lowest" value of the MAC address wins the comparison. Determining the root bridge ensures that sent frames only take one route: via the root bridge.
Certain properties must be assigned to the respective ports in order to be able to filter whether the traffic should be blocked or allowed. The following switch port designations are distinguished:
This port describes the shortest route to the root bridge. Each switch (except for the root bridge) assigns this role to exactly one port. This is a forwarding port. It forwards data packets.
These ports are all "non-root ports" that forward network traffic. So if the connection ends on a root port, what is known as a “designated port” can be found at the other end. The root bridge only has designated ports. This is also a forwarding port.
These ports are in blocking or discarding mode. They arise when two Designated Ports are connected to each other. In these connections, one side becomes the alternate and the other remains a designated port.
These ports are disabled on the switch and do not forward network traffic.
With the spanning tree protocol, the algorithm is run through until the root bridge can be determined and the ports have been configured accordingly. As long as no traffic flows. The ports go through different states during the spanning tree protocol until they have their final role:
Process | 802.1D STP | 802.1w RSTP |
Administrative off | Disabled | Disabled |
Disabled by STP BPDU are accepted | Blocking | discarding |
No "payload traffic" | Listening | discarding |
No "payload traffic" MAC table is filled | Learning | Learning |
normal function | Forwarding | Forwarding |
The algorithm is defined according to IEEE 802.1D and specifies the following port speeds:
| Speed | 802.1D-1998 (STP) Cost | 802.1D-2004 (RSTP) Cost |
| 10 Mbit / s | 100 | 2.000.000 |
| 100 Mbit / s | 19 | 200.000 |
| 1 Gbit / s | 4 | 20,000 |
| 10 Gbit / s | 2 | 2,000 |
| 100 Gbits | N/A | 200 |
In order to initially recognize in a network which switch is assuming the role of the root bridge, all switches exchange so-called "Bridge Protocol Data Unit" frames (BPDU) with one another in accordance with the Spanning Tree Protocol. This BPDU contains the bridge ID and the root ID. The root ID (consisting of 6 bits) is initially always the same as the bridge ID. As explained in the section above, the bridge ID consists of the priority and the MAC address.
In the course of the spanning tree protocol, the switches compare the received bridge IDs with one another. If these are smaller, the switch updates the received frame's root ID with its own. The switch ports will also be adjusted. This creates new root ports, designated ports and blocked ports.
The priority can vary between 0 and 61440. The most common default is 32768. Valid priority values are 0 or multiples of 4096. The lower the number, the more likely the switch is selected as the root bridge.
The priority consists of the costs (Cost) (unitless) and the VLAN ID (12 bits). So 2^12 = 4096 VLANs are possible.
With the spanning tree protocol, the algorithm is run through until the root bridge can be determined and the ports have been configured accordingly. No data will flow until then.
