Virtualization
How to scale virtualization
Enterprises should consider phased implementation for maximum benefit.
by Chris Wiborg
In times of economic uncertainty IT budgets are often trimmed. Server virtualization, though, can help businesses get more value from both the hardware and software infrastructure, and from the labor required to keep IT systems up and running.
To leverage the benefits of virtualization across the enterprise without being hindered by the increased complexity introduced by such change, IT departments should adopt a holistic approach. By planning carefully and using services available in the network, IT organizations can adopt a logical, phased implementation of virtualization technology that will provide maximum benefit to the enterprise.
Enterprises move through three major phases on their path toward virtualization. The first step is infrastructure consolidation. Existing computing and storage resources, as well as disparate network capabilities, are moved into newer and fewer devices capable of performing multiple functions. Individual tower servers, for example, are replaced by rack or blade chassis models, and individual network appliance functions are consolidated into integrated services routers. Servers, storage equipment, networking gear and other data center resources are also likely to be consolidated into a smaller number of physical locations.
In the next stage, by taking advantage of virtualization, applications can use logically partitioned computing, network and storage resources that can be rapidly provisioned and assigned. Activities undertaken during this phase often begin with a progressive deployment of hypervisor technology across different server environments: development, staging and production. Server processing resources are assigned to one or several shared computing pools. Storage resources can be collapsed into a single storage-area network (SAN) using virtual SANs (VSANs), which convert separate storage islands into a single fabric that maintains scalability and security. Storage resources are likewise combined into one or more shared storage resource pools.
A critical part of the move from consolidation to end-to-end virtualization is an examination of the underlying network infrastructure. Taking a comprehensive approach that considers how technologies work together as virtualization is implemented improves organizational alignment.
Service partitioning and aggregation virtualize the services of the network itself for use by individual virtual machine (VM) contexts. Some of these services, such as VLANs and virtual private networks (VPNs), have been an inherent part of network functionality for decades. Newer services play a variety of roles as companies embrace new application architectures.
Load balancing ensures the performance and scalability of applications by distributing inbound requests, effectively virtualizing destinations.
Network service partitioning facilitates consolidation and resource use by instantiating services such as firewalls and load balancing into multiple virtual contexts.
With switch virtualization, two or more physical switch chassis cluster together into a single, logical entity for simplified management, enhanced reliability and scalability.
VSANs increase scalability and performance through a logical grouping of Fibre Channel switch ports within or across connected devices, enabling the segmentation of SAN traffic and associated fabric services.
I/O virtualization eliminates the need for redundant physical host adapters, cables and ports through the consolidation of Ethernet and Fibre Channel I/O into a unified fabric (Fibre Channel over Ethernet).
The final step in a phased virtualization rollout is flexible service automation. This phase allows an intelligent VM-aware network fabric to rapidly and automatically detect and respond to the changing needs of applications and to provision processing, storage and security resources as needed. The cornerstones of this phase are automated service provisioning, automated security responses and self-healing systems. Improvements in reliability can be achieved by reducing complexity and the potential for human error, while a lower total cost of ownership can be realized through a reduction in manual intervention and the improved utilization of allocated processing and storage resources.
Automating these processes allows for a policy-based, adaptive infrastructure that will greatly improve business agility. Enabling resource virtualization with intelligent, application-aware networks that can respond to changing conditions and individually optimize the performance of each application allows the IT department to be more responsive to the business.
Chris Wiborg is enterprise architecture marketing manager, Cisco, San Jose, Calif.
For more information
(click here)
A phased virtualization plan includes infrastructure consolidation, a thoughtful adoption of virtualization technologies and techniques, and the leveraging of available capabilities for service automation.