Introduction

With the increasing growth of Data Centers spread around the world, the necessity of a network that supports the demands of the data center communication became really strong. For this task, the Data Center Bridging(DCB) appears as a solution. The DCB consists in a series of enhancements to IEEE 802.1 technology that defines the Ethernet pattern.

A data center holds and processes huge amounts of data and requires a network capable of dealing with the transmission of a enormous quantity data. The diversity of the stored data also introduces priorities to how these networks treats the intercommunication. With this and other desirable characteristics, a IEEE task group start to work in DCB to provide Ethernet the capability of fitting on these criteria.

The original Ethernet acts as a best effort network technology that allows package lost in certain circumstances and does not provide flow control and error correction. In the IP architecture, the flow control and error correction are usually done by the Transport layer with the TCP protocol. Although, in the data center network these tasks are desirable to be performed by lower layers, due to the particularities of data center architectures and functions.

The main proposed modification by Data Center Bridging are: Priority-based Flow Control, Congestion Notification and Enhanced Transmission Selection. The next sections of this work aims to explain and describe how all these changes works, and also some of the new modification that are being developed.