High Availability


Current IT Industry

Data center technology has developed in recent years. Rising popularity has led to many major companies focusing on public cloud spaces. However, many people with sensitive data are struggling to balance convenience and security. Data centers are now moving from traditional physical storage to VPS (virtual private server)(VM ware and hyper vision) With such changes, IT management has become significantly more complicated. Nevertheless, ensuring a stable cloud service is crucial in an age of heavy data loads and valuable information.


Since 2001, Openstack entered the cloud market and has grown rapidly, attracting attention from various fields. Openstack has recently developed into an open source provider of cloud technology. Two years ago, the newest edition of Openstack provided a means of stability and high availability.

What Vling E Business Offers

Vling E Business focuses on Openstack to build high availability. Typically, cloud services with high availability have “master standby backup” features. Hardware specifications include three cabinets with three servers (computer nodes). Three control nodes work simultaneously as a main controller (neutral). Based on these controllers, the above controllers work as a hybrid cloud kernel using open source Openstack technology. Computing nodes (NOVA) can be added into the cloud according to requirements. Traditionally, three servers can be added into the cloud to provide a NOVA service. Through configuration computing nodes can be turned into cloud storage. Nodes can be configured into CPU’s and memory for additional VPS’s. Configurations will decide the future of computing resources  i.e. (CPU, memory, storage, capacity)

Successful cloud infrastructures are configured by Openstack technology and are supported by network hardware. For example, three stackable switches are configured as one switch in order to support high demand services. Vling E Business will configure Openstack using high availability standards. Most modules are run as master, standby and backup. This provides a safety net in case of failure. Cloud VPS running on these cloud servers will have the same safety net in place.

Our cloud utilizes Checksum, CRC (cyclic redundancy check) together with a failover solution. Our Openstack cloud has been uses in hybrid clouds for one year with excellent HA performance. We setup cloud services including API, RPC, Cache, Stateful with enough mirrors of 3 or 3+ redundancy for HA performance. Pacemakers have been used in all of our main nodes including Neutron, Keystone, and Happroxy.

Vling cloud system uses Openstack’s preferred RabbitMQ and mirror queue for system control which works to guarantee a synchronized message queue. Furthermore, this system will be able to run at a healthy state despite single node failure. When a node failure is detected, Pacemaker combined with a VPS allows for on-demand cold or live migration. Three ceph monitors work in the block storage node. These nodes also store 3 copies amongst 3 servers, which are equally distributed amongst corresponding cabinets.

Our HA neutron node uses 2+ mirror DHCP service, while a VRRP (virtual router redundancy protocol) has been used for the SDN (software defined network) service. The VRRP service also comes with a network interface (Linux process) starting with ha*. This interface constantly keeps the VRRP alive with checkups every 2 seconds. Any network fails will lead to an immediate FIP operation to float external public IP’s to the standby VPS.

The Best HA Infrastructure

Despite downtime for physical servers, a cloud system will eliminate any disruption to user experience. You will no longer have the inconvenience of downtime or the need to fix any problems that occur. Traditional systems incorporate redundant designs like that of a dual power supply and RAID. Vling incorporates a design based on stackable HA cloud engine switch (core switches, 2nd level and 3rd level switch).

Each host comes with a bonding network using, load balance and LACP (link aggregation control protocol). Our HA cloud reduces hardware failure to less than half before any users are able to experience change. Our downtown Vancouver cloud infrastructure is based on a 3 cabinet system. Individual cabinets (1/3 of hardware resources) can also be powered down without any interference to user experience.

LiveZilla Live Chat Software