Designing A Scalable Workload Management System

3616 WordsOct 24, 201415 Pages
2.4.1 Neutron Overview Without including any network-specific functionality, it would be possible to develop a scalable workload management system. While the connectivity between compute nodes is required and also for the external access, it is important to have network-specific functionality, but it would be possible to benefit from the existing networking infrastructure to allocate IP addresses and relay data between nodes. However, the main problem with such an approach is the network management system would not be able to separate traffic between users securely and efficiently when there will be a multitenant environment[13]. Neutron is a part of OpenStack project focused on delivering networking as a service and manages the…show more content…
A simple model of virtual network, subnet, and port abstractions to define network resources are the base for creating OpenStack Networking component. Similar to a VLAN in the physical networking world, network in OpenStack is an isolated layer-2 segment. Subnet is a block of IP version 4 or version 6 addresses, set as an address pool from which OpenStack can assign IP addresses to virtual machines (VMs). Each subnet is stated as a Classless Inter-Domain Routing range[13]. 2.4.2 Neutron Architecture The main process of the openstack networking component is neutron-server, which transmit user requests from OpenStack Networking API to the configured plug-in. There are three agents that interact with the main neutron process though the message queue or the OpenStack Networking API[9]: - Neutron-dhcp-agent: The main function of this agent is to provide Dynamic Host Configuration Protocol (DHCP) service to all tenant networks[9]. - Neutron-l3-agent: The main function of this agent is to translate and forward Layer3/Networking address to enable external network access for VMs on the tenant networks[9]. - Neutron-*-agent: This is an optional plug-in agent, which is responsible of performing local virtual switch configuration on each hypervisor[9]. When Nova launches a VM instance, the service communicates with OpenStack Networking in order to plug each virtual network interface into a particular port. 2.4.3 How Neutron works
Open Document