Type
Linux Virtual Server (LVS) is load balancing software for Linux kernel–based operating systems. The Linux Virtual Server as an advanced load balancing solution can be used to build highly scalable and highly available network services. It is using a Layer 4 switch, and able to provide the services extended to the hundreds of thousands to tens of thousands of simultaneous connections at very high speed.
There are many different types of operating cluster in Hanbiro. Typical is three of the following:
Load Balancing (Typical cluster form)
One load-balanced server is the most common form of clusters distributes requests to multiple real servers. However, there is a disadvantage that if load-balanced server causes a failure, whole service will be stopped.
Load balancing distributes request processing across multiple servers.
High Availability and Load Balancing
Two load-balanced servers are the roles of master and standby in Failover form. If error occurs while the master server is responsible for the load balance, the standby server in standby status will be immediately taken over load balance. If the master server is restored completely, the standby server goes back to wait state. This is the most stable form.
Two load-balanced servers in the master server distributes requests of clients across multiple real servers. When the master server is down, the standby serser will replace the role of the master server. If the master server is restored, the standby server goes back to standby mode. OS on a cluster server is also available as a Windows server.
Streamline High Availability and Load Balancing (The real server is good at both High Availability and Load Balancer)
Two real servers are Cluster types for serving high availability functions. Due to the fact that the load balanced servers are not placed separately, although it can bring cost-effective, but there is a downside that Real servers are limited of Linux platform, so it does not suit for a large service.
Two real server will act as a load balancer servers with Failover feature. If Real Server 1 / the master server has problems, Real server 2 will do load balancing and Real Server 1 are excluded from the cluster group.
