In this article, the basic concept of OpenStack Octavia loadbalancers are explained. This includes the uses, options, benefits and how to create a loadbalancer.
Loadbalancers can be an essential part of a cloud infrastructure. Loadbalancers can be used to distribute the incoming traffic to multiple servers. This can in turn increase the availability of the servers and the applications. OpenStack Octavia is a loadbalancer service that provides loadbalancing services to the OpenStack cloud. Octavia is a on-demand and reliable loadbalancer service. Octavia is a replacement for the older Neutron LBaaS service.
There are two types of Loadbalancers in OpenStack Octavia which can be used to create your Loadbalancer setup.
A single loadbalancer setup is a simple setup where a single loadbalancer is used to distribute the incoming traffic to the backend servers/applications. This setup is suitable for non-critical applications because if the loadbalancer fails, the whole setup will be down.
The Active/Standby Loadbalancer setup is a more reliable setup where the standy loadbalancer will take over the traffic if the active loadbalancer fails. This setup is highly recommended and essential for critical applications. This setup can be used to make sure the availability of the servers and applications can be assured.
Loadbalancer flavors are used to chose between the Single and Active/Standby Loadbalancer setups. Most of the time, the Active/Standby Loadbalancer setup is used because of its reliability and availability. Loadbalancer flavors are also used to define the performance of the loadbalancer. The flavors can be used to define the performance of the loadbalancer like the number of connections, the number of requests, the bandwidth, vice versa. Based on your requirements, you can choose the flavor that suits your needs. Most providers indicate the performance of the loadbalancer in the flavor description.
Listeners are used to define the incoming traffic to the loadbalancer. The listeners are used to define the protocol, port, and the pool to which the traffic should be forwarded.
A pool is a group of backend servers or applications to which the traffic should be forwarded. The pools are used to define the protocol, the algorithm, and the backend servers for the specific Listener.
Pool algorithms are used to define the way the traffic should be distributed to the backend servers.
Session persistence is used to make sure that the traffic from the same client is always forwarded to the same backend server. This is used to make sure that the session data is always available on the same server. This is essential for applications that require session data to be available on the same server. OpenStack Octavia currently supports the following Session persistence methods:
Monitors are used to define the health of the backend servers. The monitors are used to define the protocol, the interval, the timeout, the retries, and the status of the backend servers. Whenever a backend server or application is down, the monitor will mark the server as down and the traffic will not be forwarded to that server anymore. This way you can make sure that the traffic is only forwarded to the healthy servers to avoid downtime for the users.
SSL Termination is used to decrypt the incoming traffic and forward it to the
backend servers. This is used to offload the SSL encryption from the backend
servers. This way the backend servers can focus on the application and the
loadbalancer can focus on the SSL encryption. The storage of the SSL
certificates is be done with OpenStack Barbican (Secret Manager) and can be
added to the loadbalancers on creation or at a later point. It is required to
use the TERMINATED_HTTPS listener protocol to enable SSL Termination at the
Loadbalancer.
Chosing the best location for you loadbalancer can be an essential part of the setup. We recommend checking the geographical location of the backend servers to make sure the loadbalancer is placed in the desired region/availability zone (AZ).
Loadbalancers can be used to forward external traffic to the backend servers but they may also be used to forward internal traffic within your private network. This is something to keep in mind when desiging your cloud infrastructure.
When creating a loadbalancer setup, it is recommended to use the Active/Standby Loadbalancer setup. This is because of the reliability and availability of the setup. The Active/Standby Loadbalancer are setup to make sure the availability of the servers and applications can be assured whenever one of the loadbalancers fails.
When you create a loadbalancer we recommend to use an Floating IP address for the loadbalancer. This way we can always swap the IP between loadbalancers in case you want to upgrade or replace the loadbalancer.
Loadbalancers can be an essential part of a cloud infrastructure. Loadbalancers can be used to distribute the incoming traffic to multiple servers. This can in turn increase the availability of the servers and the applications. OpenStack Octavia provides a on-demand and reliable loadbalancer service which is easy to use and manage for your cloud infrastructure.
Now that you know everything you need to know about loadbalancers, you can start creating your own loadbalancer setup. We highly recommend to read the article about creating a loadbalancer with webservers to get started with OpenStack Octavia.