Major American Retailer Turns to Cirba to Gain Control over AIX in PowerVM® Environments
The Challenge: Platform Complexity Makes Optimization Difficult
Key to driving up efficiency in PowerVM environments is the ability to place LPAR's on frames in a way that meets their business needs and at the same time makes optimal use of overall capacity. Like most PowerVM users, this was a challenge for this large retailer, and without dedicated analytics, the complexities of the PowerVM platform made it difficult to optimize placements. Specifically, it was not possible to take into account hardware capabilities, workload demand patterns, shared processor pooling, entitlement management, VIO Server configurations, and other factors when making decisions using existing tools and processes. This led to sub-optimal LPAR placement, which adversely impacted density and made it impossible to reclaim stranded capacity. It also made forward-looking predictive analysis impossible, which affected the responsiveness and overall agility when it came to new business needs.
The Solution: Control for LPAR Placement, Infrastructure Sizing and Operations
Cirba was chosen by the retailer to gain precise control over the PowerVM environments. Cirba's support for advanced PowerVM and AIX features allowed the complexity of managing these environments to be tamed, while at the same time providing a common management paradigm across other platforms as well. Cirba's Control Console provides a clear visualization of the capacity of LPARs, frames and overall environments, and generates precise actions, based on advanced policies, to keep everything operating "in the green". Also, Cirba's capacity bookings feature, and the new Reservation Console, were seen as a way to operate in a more proactive and predictive way by gaining visibility into new workload requirements and understanding if and where they fit.
Cirba currently provides control for over 5000 LPARs in this retailers environment. It is actively used to control capacity, generating LPAR rebalancing instructions, as well as capacity reclaim recommendations (e.g. down-sizing LPARs). It is also used to manage Shared Processor Pool (SPP) sizing and membership, Virtual I/O Server operational requirements (via specialized VIOS policies), and enforce the capacity requirements for PowerHA pairs.
The investment in Cirba has paid off. The more precise management of placements and automated analysis of reclamation opportunities has led to over $1M in hardware cost avoidance in the first 12 months of use.
The Impact: Cost Control and Agile, Automated Operations
Cirba has enabled the retailer to successfully pursue its goals of increasing density and freezing hardware spend over the next three years in its virtualized AIX infrastructure, resulting in significant savings. The company has also seen operational benefits and has been able to increase automation which is an important aspect of its long term vision.