ServerMotion offers two types of servers: dedicated servers on dedicated hardware, and virtual cloud servers deployed on our high performance virtualized cluster. Both have their pros and cons, and any wise systems administrator will only make the decision after being well informed.
Dedicated servers offer access to dedicated hardware resources. The full capacity of the hardware is available at your disposal, and the hardware is not shared with other users at anytime. Although dedicated resources offer the absolute power of the hardware, they also succumb to the faults of possible hardware failure. If the hardware of a single server fails, the entire server will be offline until it is repaired. Additionally, if the server does not have dual power supplies spread across A/B power feeds, the server can go down due to a single power circuit failure in the datacenter.
Cloud computing offers access to a virtualized server that deployed on a cluster of hardware nodes. The resources of the hardware are shared with other virtual servers, and unless deployed on a private cloud, the hardware is shared with other users. However, the virtualization layer provides for even allocation of hardware resources and is able to guarantee a certain level of performance.
In the case of ServerMotion, this guaranteed level of performance is allocated in the form of Cloud Nodes. Each node represents a collection of resources — RAM, CPU, Disk and Bandwidth. By assigning multiple nodes to a given virtual server, it is possible to size the server’s resources to your exact needs, minimizing cost and waste.
Cloud solutions are deployed across multiple power circuits and multiple hardware nodes, allowing for any single hardware failure to occur transparently to the virtual server. In the event of a hardware failure, the server is simply moved to another hardware node in real-time.
Since the servers themselves are virtual, it is also possible to quickly enlarge or shrink the underlying resources available to a server. In the event of high load or increased capacity needs, a server can be quickly resized to have access to more resources of the underlying hardware.
Both dedicated servers and cloud servers still must be implemented in a fault tolerant way. By provisioning dedicated and/or cloud servers in multiple datacenters and using technologies such as Round Robin DNS, Anycast DNS with rapid failover, and load balancers or heartbeat within the datacenter, it is possible to build a solution that has absolute 100% uptime.
