Bandwidth control in multi-tenant virtual networks
Machines, systems and methods for managing quality of service (QoS) in a multi-tenant virtualized computing environment, the method comprising: collecting transmission rate statistics associated with data communicated in a virtual network, wherein at least one virtual switch monitors communications initiated by at least a virtual machine (VM) executed on a host machine serviced by the virtual switch; determining, by way of the virtual switch, profile parameters associated with a first communication initiated by the VM belonging to at least a first group, wherein a connection request is submitted by the virtual switch to a traffic controller to assist in establishing the first communication; classifying the connection request for establishing the first communication according to the profile parameters associated with the first communication; determining a first aggregated transmission rate associated with the VM that initiated the first communication based on the classifying.
1. A method of enforcing a service level agreement (SLA) in a multi-tenant communications network, the method comprising:
assigning a first QoS level to a first communication initiated by a VM belonging to at least a first group, wherein a connection request is submitted by a virtual switch to a traffic controller to assist in establishing the first communication,
wherein a first SLA associated with the first communication defines acceptable throughput and latency metrics for the first communication based on profile parameters associated with the communication, wherein the first SLA further defines a throughput metric comprising aggregated data transmission rate between a first VM belonging to a first policy group and a second VM belonging to a second policy group, wherein the first SLA further defines at least two or more levels of delivery priorities for the first communication based on profile parameters associated with the communication, and wherein the first SLA further defines a latency metrics that is enforced over traffic based on a strict reservation of bandwidth over potential paths between a first VM in a first group and a second VM in a second group;
wherein the traffic controller determines whether the connection request can be submitted from the virtual switch to a physical switch;
wherein the first SLA defines a throughput metric and a latency metric that is enforced by provisioning of guaranteed resources and enforcement by an underlying switching infrastructure;
wherein the underlying switching infrastructure comprises one or more virtual switches that shape traffic and set a service class identifier within a packet header of the first communication;
wherein the underlying switching infrastructure comprises one or more physical switches that operate based on instructions issued by a centralized controller; and
wherein the first SLA further defines at least two or more levels of delivery priorities for the first communication based on profile parameters associated with the communication.