Scalability

One of the most important things that applications need to be able to do in the cloud is scale up and down rapidly to meet demand. Auto-scaling is a unique feature of cloud computing and one of Virtual Ark’s core competencies. Virtual Ark leverages its own intellectual property and tools like Right Scale to set thresholds based on the traffic forecast. When a threshold is exceeded, Virtual Ark will spin-up a new instance of the webserver, and roll it into the load balancer pool automatically. Once traffic falls below the minimum threshold, Virtual Ark will remove one or more servers out of the pool. Virtual Ark monitors this activity and runs load and stress tests to ensure more infrastructure can be added at exactly the right time. Virtual Ark ensures that the application behaves as it should, and no anomalous cases will cause servers to spin up unnecessarily. This allows Virtual Ark to use (and pay for) only the infrastructure that is necessary when and while it is needed.

Virtual Ark also scales the database layer horizontally to accommodate database scaling. There are several architectures available including a master-master active passive cluster, also known as circular replication. In this configuration, the database will send all completed transactions to the other server in the cluster. The passive server can also handle read traffic for the application. This allows Virtual Ark to not only achieve high availability, but also scalability by sending much of the traffic off to another server. As the load grows further, Virtual Ark will add an additional read-only slave into the configuration. It can then further scale out reads. As Virtual Ark does this, all writes will continue to go to the single active master database. If the load on the single master database continues to be a problem, Virtual Ark may vertically scale that node in place. This is accomplished by spinning up a new larger server instance. This scaled up will be delivered in place without down time or disruption to the application.

Another way that Virtual Ark achieves scalability is through the storage architecture. Virtual Ark configures the storage subsystem to provide the reliability obtained from a fully redundant RAID storage system. Technologies such as snapshots allow Virtual Ark to recover from many types of failures. In addition, it is possible to increase the I/O transaction rate by mounting multiple volumes on one instance and striping file systems across them. This can increase the performance of random workloads as more heads can be seeking at a time.