<-- Back to full color view

Getting the Most from Your Virtual Infrastructure

Originally published June 15, 2009

The current economic cycle has placed intense pressure on IT budgets. CIOs are being asked to consolidate IT operations and cut costs – without reducing their service levels. As a result, surveys show that CIOs are making server virtualization and consolidation projects their priority to optimize IT resources and services. 

Most enterprises have already migrated to storage area networking for a flexible, lower-cost, higher-service-level storage infrastructure. Currently, thousands of enterprises are adopting server virtualization technologies that provide the same benefits for the rest of the IT infrastructure. This synergistic combination of storage, networking and computing virtualization has created a new category of infrastructure software called virtual infrastructure.

Virtual infrastructure is rapidly transforming the IT landscape and fundamentally changing the way that people compute. Today’s powerful x86 computer hardware was designed to run a single operating system and a single application. This leaves most machines greatly underutilized. Virtualization lets users run multiple virtual machines on a single physical machine, sharing the resources of that single computer across multiple environments. Different virtual machines can run different operating systems and multiple applications on the same physical computer.

Virtual infrastructure gives administrators the advantage of managing pooled resources across the enterprise, allowing IT managers to be more responsive to dynamic organizational needs and to better leverage infrastructure investments. It provides a flexible way to rapidly provision IT resources to match high-availability requirements, improve utilization from existing IT resources and handle increased workloads and application response times. It also helps drive operational efficiency with consolidation of underutilized IT resources and reduced hardware requirements.

But the value of virtualization can’t be fully realized unless IT groups can track and analyze data from their overall IT resources quickly and accurately. They also need to measure virtualization successes in terms of value to business units – whether it is calculating reduced customer wait times or total cost per help desk contact per month.

CIOs are under pressure to reduce capital expenditures and want to move away from past practices that led to server overcapacity and sprawl. For server virtualization projects to yield positive results, data center professionals need to have visibility into the use, availability and future performance of both physical and virtual infrastructure resources. Some key questions to consider:

  • How should you plan, manage and deploy IT resources to reflect individual, departmental and enterprise priorities and constraints?

  • How can you find a safe, manageable and logical way to virtualize your physical IT resources and conduct server-consolidation exercises?

  • How can you forecast utilization, rebalance workloads and proactively plan capacity to avoid server sprawl?

  • How should you charge back to business units?

  • What operational bottlenecks are affecting IT productivity?
Tracking the answers to these questions requires a system that can easily collect data from heterogeneous environments, perform analysis to determine the appropriate virtualization strategy given the constraints, and predict needs and outcomes to quickly make changes to service-level agreements and costing models.  

A comprehensive solution should:
  • Access, process and aggregate IT performance data from all types of physical and virtual data sources.

  • Provide a complete picture on utilization, and availability of resources IT organization has at its disposal.

  • Identify and monitor over- and underutilization of IT resources as a function of current capacity, planned usage and growth against costs.

  • Predict various scenarios for the deployment of virtualized resources to match business requirements. Apply analytics for optimal placement of virtual resources and physical-to-virtual server migration. 

  • Link usage and availability of IT resources with costs for proper budgeting and planning.

  • Use different methods to allocate and charge costs back to the business units that consume resources.

Analyze and Optimize Your IT Resources

Resource optimization models should analyze both physical and virtual IT resource needs. A company needs to bring the virtual data into the IT data mart and correlate it with business unit information to expand or broaden the picture of how virtualized environments will play in IT infrastructure. A critical first step is to model the physical and virtual infrastructure to find problems before they affect the business. To accomplish this you need strong data integration capabilities to extract, transform and combine performance data from virtual data sources with volumes of data coming from other physical IT sources. Pre-defined data models, adapters for third-party data sources, and a library of prepackaged and custom data transformations will enrich the IT data mart with statistically relevant, normalized and timely information for further analysis. 

A range of analysis and reporting techniques or capabilities is critical, each oriented toward optimizing IT resource, financial and service-level strategies (e.g., workload profiling, resource consolidation, service-level measurements and improvements, capacity planning, chargeback). The capabilities should be able to meet the needs of different types of IT users, including senior IT management. Analysis should cover everything from standard reports (such as measuring CPU, memory, and disk and network utilization) to forecasting scenario reports that provide insights into future planning needs for demand, capacity and licensing. The results should be displayed in a dashboard to determine what actions can be taken. The following questions can be answered by employing ad hoc, interactive or batch reporting and analytical capabilities in IT operations:
  • Where are over-provisioned or under-provisioned virtual machines?

  • How can you reconfigure server capacity quickly during failure or disaster events?

  • How many new virtual machines are needed to improve performance and meet service level agreements?

  • How many virtual machines can we run per host before performance degrades? Are virtual machines provisioned properly with the right amount of memory to avoid degradation of application run times and performance?

  • How will techniques like memory ballooning and page swapping affect active host(s) in terms of memory capacity and ability to accommodate processing load from failed or maxed out host(s)?

  • Can you forecast virtual machine CPU and memory growth trends and forecast to match with projected workloads from business units?  

  • What is a healthy ratio of virtual machines per physical server host to prevent the over-allocation of infrastructure resources that contributes toward virtual sprawl?

  • Can we continue to suggest further optimizations in hosting and migrating applications and virtual machines based on monitoring application footprints continuously?
Properly assessing physical server and application candidates for virtualization can help achieve operational flexibility and cost-effectiveness and improve hardware utilization. Potential cost savings can be realized when users address the biggest and most obvious inefficiencies by grouping similar workloads, availability needs, security and agility to hosted virtualized platforms. Analytic algorithms and what-if simulations can guide IT staff to conduct scenario planning exercises for optimum workload placement across heterogeneous systems. With an analytical approach to virtualization, IT departments avoid being trapped into imbalance across the IT resource footprint. And the same analytics-driven approach will help to link IT metrics to corresponding business metrics and help senior IT management staff to demonstrate business value. Pointing out how much less time it takes to process a mortgage application or insurance claim resonates with business units in a way that “We’re using our servers more efficiently” just doesn’t. Defining IT goals in business terms and measuring value from IT investments that business units can understand is critical to prevent IT budget cuts that will undermine the entire enterprise.

SOURCE: Getting the Most from Your Virtual Infrastructure

  • Tapan Patel
    Tapan Patel is Global Marketing Manager with SAS, market leader in business analytics software and services. He focuses on persona-based marketing, awareness and messaging for data mining, model management, grid/HPC computing, in-database processing, and IT management software. Prior to his role in product marketing, Tapan spent many years working with customers, partners and industry experts as a product manager, market research analyst and technical sales engineer. He may be reached at tapan.patel@sas.com.
  • Bill Wright
    Bill Wright is Sr. Manager Global Alliances for VMware, responsible for the business intelligence application segment. Before VMware, he held business development and global account management positions with ATG, IDG, Hitachi Data Systems and Amdahl. He may be reached at Bwright@vmware.com.


Want to post a comment? Login or become a member today!

Be the first to comment!


Copyright 2004 — 2020. Powell Media, LLC. All rights reserved.
BeyeNETWORK™ is a trademark of Powell Media, LLC