Operational Risk Analysis

CiRBA provides comprehensive risk analysis, whether analyzing environments for transformations, such as through candidate qualification risk analysis, or workload contention probability analysis. From an operational perspective, CiRBA provides the ability to mitigate risk through setting thresholds for alerts to defining risk tolerance levels suitable for workloads when rebalancing an environment. CiRBA also enables organizations to model and plan for distinct risk scenarios, such as cabinet failures, configuration risks, and the analysis of contention probability within shared environments, to ensure that performance risks are minimized.

Planning for Potential Failures

CiRBA enables organizations to model and plan for potential failure scenarios. CiRBA combines rebalancing analysis with rules to simulate certain types of failures to assess whether sufficient capacity remains in the surviving servers to host the affected workloads. For example, if a VMware cluster is spread across multiple cabinets, then analysis scenarios can be configured to simulate the failure of each individual cabinet. If sufficient capacity remains in the surviving cabinets for the affected workloads to be restarted, then the scenario is considered successful. CiRBA also supports more advanced analysis that combines scenario planning with forward-looking trending and rules to restart only top-tier applications, in order to provide a very high degree of resilience. Critical to this analysis is CiRBA’s rule-driven modeling of failure scenarios to determine whether sufficient spare capacity exists, and whether that capacity is configured properly to survive failure (rack distribution, etc).

CiRBA supports analysis of failures at multiple levels to represent various DR and HA scenarios:

Uncovering Potential Configuration Risks

The proper configuration of IT infrastructure is critical to many aspects of operation, from maximizing reliability, to lowering risk during routine administrative activity, to enabling automation across large numbers of servers. In virtual environments, configuration management is even more critical, as the capacity allocated to workloads is also configurable. Configuration is one of the three pillars of CiRBA analysis, and is a critical element of every transformation plan, and in the ongoing management of physical and virtual infrastructure. CiRBA’s rule-based risk analysis identifies areas where technologies may not be compatible, where sufficient resource may not be allocated, and where remediation is required in advance of a planned change. CiRBA also provides configuration drift and divergence analysis against a gold standard, peers, or different environments (such as production vs. DR or staging vs. production), to catch potential issues before they become problematic.

Predicting Potential Contention for Resources

A more advanced form of risk analysis that is particularly suited to shared and virtual environments is contention probability analysis. CiRBA enables the analysis of operational patterns and statistical characteristics of running workloads in order to determine the risk that workloads will compete for resources, and therefore experience performance issues. CiRBA enables organizations to fine-tune analyses and density rates according to the desired probability that they will (or will not) contend for resources. CiRBA enables organizations to design infrastructure and determine workload placements according to different risk tolerances. More critical workloads can be placed together, and the environment sized, so that the workloads housed within it have very little probability of competing for resources. Less expensive infrastructure, or more densely populated infrastructure, can be established for less critical workloads. The ability to statistically determine the risk level within which an environment is running (even if problems have yet to be experienced) is key to the proper management of resource allocations and VM densities.