|
Virtualization is a key technology that is changing the way system hardware is used. It is also changing the way desktop environments are delivered, and how users interact with core systems and applications. Outcomes from virtualization include improved system utilization, reduced downtime, improved business continuity, and better service level achievement.
However, there are significant inhibitors to a successful virtualization deployment. Specifically, the management and monitoring software that comes with the common virtualization platforms is not enough to meet the business or technical goals of virtualization. It does not provide the cross-domain metrics for technical administrators to effectively monitor and manage performance on a wide scale. It does not provide the correlation across physical and virtual systems that IT service managers need to ensure fast problem triage and resolution. It does not provide the high-level performance and service measurements that CIOs need to assure business service levels.
Virtual environments instead need comprehensive and sophisticated performance moni-toring tools that understand the whole environment – physical and virtual – and that can track, analyze, and prevent problems across the end-to-end IT service infrastructure. Such solutions need to span technologies and platforms across diverse multi-tier environments in order to deliver business goals for the CIO, service goals for the IT service manager, and technical goals for the system administrators. They need to cover key areas including discov-ery, physical infrastructure monitoring, virtual infrastructure monitor-ing, system and application monitoring, and connecting the pieces in a holistic view.
This EMA white paper investigates these issues in more detail, out-lining the key drivers and outcomes of virtualization, explaining criti-cal inhibitors to virtualization success (and how to overcome them), and describing core virtualization performance management solu-tion requirements. It also briefly describes one such solution, the eG Enterprise Suite, and includes a short case study highlighting this solu-tion in action.
Virtualization – A Short Introduction
Virtualization is a technique for hiding the physical characteristics of computing resources from the way in which other systems, applications, or end users interact with those resources. Server virtualization, for example, allows multiple operating environments (‘guests’) to run on a single physical server (‘host’). Remote desktop virtualization ex-tends these virtual environments to end users, allowing them to access a centrally hosted ‘virtual desktop’ from any physically remote system.
EMA research reveals that over 90% of enterprises rate certain business drivers as either critical or important to their decision to implement virtualization. These include improved system utilization, reduced downtime, improved business continuity, and better service level achievement. Virtualization helps to deliver these goals in several ways, for example:
Dynamic workload management and centrally hosted desktops increase server utilization and reduce physical desktop system requirements, while extending the usable life of existing desktop PCs.
Faster (and cheaper) support of systems, applications, and even desktops from a central location reduces repair times, delivers new services and applications faster, and improves business continuity.
Central management, support, and protection of corporate data and systems, and fast migration of failing server and desktop workloads reduces downtime due to system failure and data loss.
Centralized control allows fast response to problems and ensures data security to meet corporate SLAs and compliance objectives.
Inhibitors to Virtualization Success – and How to Overcome Them
Difficulties in Monitoring Virtual System Performance
Value-added management tools from virtualization vendors, such as VMware’s VirtualCenter, are invaluable for silo-based management of virtual systems, providing important capabilities for administration, failover, configuration, etc., within the virtual environment. However, they are not designed to be complete performance monitoring solutions, so in isolation they are not sufficient to address broader and more complex performance and management problems. Without sophisticated solutions designed specifically for performance management:
Technical administrators waste time and effort manually correlating, diagnosing, fixing, and reporting on performance issues in larger deployments.
IT service managers have limited visibility with no view of performance across multiple virtual and physical systems.
CIOs cannot guarantee service levels or meet business objectives as they have no visibility into the impact virtualization has on business services.
A substantial contributor to these problems is that these tools are unaware of anything outside their limited domain. They do not ‘see’ the physical environment, unallocated capacity outside their own resource pool, external constraints like net-work and storage limitations, or the complex resource utilization across tens, hundreds, or even thousands of hosts and guests in large virtual infrastructure deployments. In large virtual machine (VM) deployments with multiple domains, the narrow focus on a limited set of virtual machines is not capable of seeing the bigger picture. When tracking complex user-level resource requirements, the machine-level focus is unable to see specific users and their resource utilization patterns. They also do not take into account the higher-level business needs, as they are unable to track and correlate business services and applications across multiple systems – especially when they are in constant motion due to live migration or workload orchestration.
|
