Today's complex computing environments typically result in equally complex storage infrastructures. Implementing a storage virtualisation strategy to move from physical to virtual servers often confuses the two basic questions regarding storage:
- Is it efficient?
- How well is it performing?
To answer these questions and lay the ground work for a successful virtualisation project, we recommend undertaking a storage virtualisation assessment (SVA). This is a high-level consultative and diagnostic engagement aimed at developing a profile of your current storage infrastructure.
Through a series of structured interviews, data collection, and observation techniques your storage specialists document your current storage environment and provide tactical and strategic recommendations for implementing virtualised storage.
Step 1: Discover
This phase is to develop a site specific engagement to detail the scope and scale of your storage virtualisation needs. It comprises the following:
- Storage hardware: a comprehensive review including documenting and reviewing end to-end hardware architecture, assessing the need for additional hardware and retiring of surplus infrastructure
- Capacity requirements: a review of the application requirements for the current project environment load and licensing requirements. If this will be sharing existing storage, review the existing storage capacity, trapped storage and over-provisioned LUNs. If the environment is new, scoping the required capacity with your application and server architecture specialist
- SAN topology: review the existing data flow and identify any bottlenecks within the current infrastructure
- Replication: capture information about the current bandwidth, utilisation of the pipe and bandwidth back to the storage environment to get a clear picture of the requirements to implement any new environment
Step 2: Develop the plan
Based upon the scope and scale of the assessment determined in Step 1, the next phase is to create a Storage Assessment Project Plan that details the timeline required for the measurement and review process as well as any setup required for performance monitoring purposes.
Step 3: Deploy
Your specialists can now deploy the assessment plan on site gathering and analysing a set of performance and utilisation metrics using established industry practices.
We would recommend performing the assessment during a peak load time such as end of month or end of quarter. This will give a more accurate view of the peak and average load times of the highest utilised times.
Diagnostic tools can be inserted non-disruptively to collect data. The data gathered is expressed in one of three forms; IOps (Ins/Outs per second), Mbps (Throughput) and Latency. For this project there are a number of key issues to address:
- Do any bottlenecks exist that are creating inefficiencies in getting data to/ from the storage array? There is always a bottleneck and in a perfect world each cog in the wheel (server, HBA, Switch, etc.) will have the same bottleneck. Limiting the performance impact of the bottleneck may be required by defining and remediation, changing HBA settings, tuning switch environment or tuning the storage configuration
- Replication Bandwidth: Can we send more data through the pipe than the pipe can handle? How do we prioritise the data? What are the minimum requirements to support the project?
- HBA Readiness: no single port servers
- Switch pathing anomalies
- Switch port/blade requirements
- Existing capacity utilisation and capacity required
- Load and licensing assessment
- Bandwidth to disaster recovery site assessment
The results of the site assessment and performance analysis should be documented in an Assessment Report and presented to IT management. This report should include the assessment results, performance measurement summaries and both strategic and tactical recommendations for implementing the new storage virtualisation infrastructure.