Kleber Reliability
DataTools aims to have Kleber available 100% of the time. Therefore we have put in place a number of processes to ensure this occurs.
Firstly, by utilising Amazon Web Services (AWS), DataTools can provide clients with the most reliable, robust and scalable cloud solution available.
In addition – DataTools has also implemented it’s own internal monitoring and diagnostics of Kleber which allows us to keep a close eye on Kleber traffic and anticipate any issues. This combined with our system to roll out updates with no downtime means that clients can be assured that Kleber will be available when they need it.
Kleber in the cloud – how does it work?
The Kleber servers at AWS are housed on Shared Server clusters. These clusters currently contain a minimum of 20 servers in the Sydney Amazon data centre.
For increased reliability the servers are evenly split over 2 isolated Availability Zones at the centre connected through low-latency links. This means that in the unlikely event of a service interruption at one zone – the other will not be affected.
Intelligent Load Balancing Cluster Control
The Intelligent Load Balancing Cluster Control automatically distributes and balances the incoming application traffic among all the servers running, improving the availability and scalability of the Kleber service.
The Controller also makes it possible for DataTools to automatically add new servers to the cluster when the need to increase capacity is required. These new servers are up in running in just a few minutes and when the load lessen, can just as easily be spun down.
This provides Kleber with great flexibility and the ability to cope with short or long term load increases quickly and easily.
Scalability and Auto Scaling
DataTools constantly monitors server traffic and performance to ensure the base cluster size is adequately handling the expected performance requirements with power in reserve.
In the event of average CPU usage across the cluster rising above 70% for more than 1 minute, the cluster will automatically launch additional servers, growing the cluster capacity by an additional 50%.
The cluster’s average CPU usage will be continually monitored at 5 minutes intervals. If the average CPU usage remains greater than 70% additional servers will again be automatically added to the cluster to raise its capacity by another 50%. This process will continue until the cluster’s average CPU usage over a 5 minute period drops to below 50%, at which time a single server will be removed from the cluster (while maintaining an average CPU usage below 50%) until the cluster returns to its base cluster size.
Auto Scaling enables the Kleber service to scale and handle instant changes in requirements or spikes in popularity reducing the need to perfectly forecast traffic.
Kleber Health Checks
In addition to the above DataTools conducts a health check every 5 seconds of the servers and their internal services. Should the system find a server that isn’t responding properly it will automatically be moved offline and a new server spun up in it’s place. The offline server will be monitored until it responds properly. Once this occurs it is moved back online and the extra server spun down.
This means that all servers are ready and able to accept requests when sent from Kleber at all times.
Data security in the cloud
Keeping on top of your customer’s data security can be a nightmare and the last thing you want is for it to reside on additional servers especially external ones.
DataTools Kleber makes this easy as no customer data or transaction data is ever stored on the servers. For added security, all data is transported over the internet using SSL encryption and all servers are located in Australian data centres.
