API adds VM placement brain to Cloud management platforms

CiRBA has announced a new API that enables organisations to connect their cloud management platforms to CiRBA in order to optimise new workload placements within internal clouds. CiRBA’s analytics determine the optimal placement for VMs within cloud infrastructure, both at the environment and server level, reducing the risk of capacity shortfalls and driving up VM density by an average of 48 percent. The new API also provides access to CiRBA’s bookings functionality, allowing users to reserve capacity for future needs using existing self-service portals.

Many organisations building internal clouds are looking to rely on cloud management platforms such as OpenStack, but face a challenge in bringing together all of the required capabilities. Cloud management platforms are designed to provision VMs, but do not have the ability to analyse capacity in order to determine the best environment to host a workload in or the best host within an environment to start an instance on. As a result, hot-spots and imbalances in resource utilisation will occur in the infrastructure, creating both performance issues and inefficient use of capacity.

CiRBA’s new workload routing API enables cloud management platforms to send placement requests to CiRBA, and to receive an answer back that contains the best possible environment and host-level placement for a new workload. This answer is based on CiRBA’s industry leading analytics, which consider a broad set of factors including utilisation patterns, licensing requirements, capacity availability, policy constraints and technical considerations. This brings a new level of automation, enabling cloud management platforms to dynamically leverage CiRBA’s analytics in order to intelligently process workload placement requests, which is often one of the biggest gaps in internal cloud implementations. Integrating CiRBA into this process ensures high efficiency while reducing operational risks, allowing more instances to fit into each environment while at the same time making existing instances work better. This new capability complements CiRBA’s standard control capabilities, which continuously “auto-corrects” cloud infrastructure through ongoing rebalancing and instance right-sizing.

The new API also enables capacity reservations to be made for new VMs through CiRBA’s Bookings Management System. This ensures the capacity is held for that workload when it is ready to be deployed if it is not required immediately.

“We see increased focus on maturing internal cloud operations in our customers,” said Andrew Hillier, CiRBA CTO and co-founder. “For self-service requests, there is a dire need for more intelligence in determining where these workloads go, and how much resources must be assigned to them. As internal cloud implementations scale and there are multiple environments, SLA levels or internal customers, it becomes unworkable to continue to place workloads based on simplistic or random algorithms.”

Continues Hillier: “Even more importantly, in enterprise environments, there is the recognition that instantaneous provisioning may not be as important as the ability to reserve capacity. Both are forms of self-service, but the ability for users to reserve capacity in advance is much more consistent with the way these enterprises work, where last-minute, unplanned use of capacity is the exception, not the norm.”