In the Ethernet infrastructure the STP achieves stable network, but it does not utilize the available physical connectivity in the best way, as some of the links between the same devices are put in blocking state, as to avoid network loops. And as it was already discussed, the goal of the STP is to achieve such a stable, loop-free environment for the networks to be operational. By design the goal for better utilization of all the available physical links is not part of STP. With the development of the technologies, in the data center arose the need of more and more bandwith. The virtualization alloed for more virtual machines (VMs) to run on top of a single physical server, and as each VM provides an isolated environment to run an application, or an application framework, by mimicking a server, then there were more applications sharing the physical resources of a server. In the same time the physical server’s connectivity to the network infrastructure wass the pipe used for the communication of all the applications running on the multiple VMs on top of it. And this bandwidth was shared at the level of the access layer switch to which it is connected with other servers, and all of them had to share the uplink bandwith of the switch. So, the interconnectivity between the access, aggregation and core layers in the data center infrastructures becam the real bottleneck for the communication needs of the applications running in the data center. To solve this challenge, there were several directions in which the technologies needed to develop:
- Hardware improvements, or interfaces with higher bandwiths – the technology went from 10/100 Mbps to 1 Gbps, to 40 Gbps, and now interfaces operating at 100 and 400 Gbps are available.
- Optimized software-defined architectures – the need to use more flexible designs, such as spine-leaf architecture with the ability to redefine the connectivity using programmatic approach.
- Overcoming the limitations of the communications’ protocols – the STP blocks the communication over redundant links, which means that these are capacities which are not utilized. The port channel technoclogy overcomes this limitation of the STP.
A port channel is an aggregation of multiple physical interfaces to create one logical interface. Port channels provide three important benefits:
Redundancy: If a member port fails, traffic previously carried over the failed link switches to the remaining member ports within the port channel.
Bandwidth: Traffic is load-balanced across the links within the port channel members to increase the bandwidth.
Spanning tree: Port channels are seen as a single switchport by the Spanning Tree Protocol; all physical interfaces are in a forwarding state.
Leave a Reply