Server Consolidation
and Server Virtualization
in Canada

The Power is Yours

Virtualization may appear like at first like a somewhat esoteric technical dreamland, but in fact, it is a game-changing approach that fundamentally changes the way business approaches investment in technology.

Virtualization, in Summary

Virtualization is a new approach to IT that largely liberates IT services from the traditional limitations of physical hardware. By "virtual," we mean "acts like" or "thinks it is" - thus, a virtual hard disk "acts like" a hard disk, and a "virtual server" "thinks it is" an actual computer, complete with a power supply, circuitry, disk drives and so on.

In reality, these "virtual" devices are just emulators of the real thing, unconstrained by physical properties such as size, location and speed. The platforms that provide virtual devices offer new forms of reliability, scalability and flexibility that are possible with physical devices. For example, processors that can be accelerated, hard disks that can be expanded, and servers that can be copied as needed.

There is Still a Physical Layer, but it's Better

Physical resources such as processors, memory and mass storage are pooled and then allocated to the virtual servers that need them, as required. Naturally, one is still limited to the total available processing power, but it can be re-allocated as required and with little trouble at all. Thus, if the total processing power, the total memory and the total mass storage available is enough to satisfy the total demand, no additional hardware resources are needed.

Virtual Storage

While hard disks are prone to failures and data loss, business have long used redundant arrays of independent drives to keep files online during such failures. Network Attached Storage (NAS) and Storage Area Networks (SAS) are further examples of pooling storage media into highly reliable sub-systems that can keep on working even when parts of the system are out of service.

Virtual drives are created from such a storage pool and assigned to servers that need storage. Servers themselves need not include any physical drives of their own, a savings that partially offsets the cost of an advanced storage solution. Storage resources are allocated very efficiently, since a boot drive for a server may require only modest capacity for a pristine copy of the operating system and nothing else, and administrators no longer need to provide extra capacity for future storage requirements on a server by server, application by application basis. Any virtual disk drive can be expanded at any time without complication.

Rather than managing an assortment of disparate servers with varying performance, free capacity and growth curves, the the IT manager can focus on the total requirement for reliable storage and the most cost-effective approach to meeting the demand.

Virtual Machines and Server Consolidation

Virtualization is step closer to an ideal world, in which the IT manager is not constrained by the number of available machines with which to follow the best day-to-day practices that keep the business systems online at all times. Many IT managers today recognize that server sprawl - rows of machines dedicated to this and that - bring management problems, not to mention heat and power costs, that can be very effectively addressed by consolidating all of that hardware on one robust platform.

The platform itself consists of one or more physical systems running a brand of Hypervisor. The Hypervisor is a software layer between the hardware and traditional operating systems. It's not a new idea. In fact, it's been the way mainframe computers and scientific supercomputers have run for decades. The physical platform is a Host for virtual machines - as many as the platform can handle at once.

A mainframe can typically host a few thousand simulated computers, each with it's own operating system and doing its own thing, just like a "real" computer. This mature technology has trickled down to the small-medium business level. The virtual approach provides servers and processing power on demand, the ability to maintain production when parts of the system are broken, and disaster recovery options that are just not feasible with rows of physical servers.

Typically, numerous legacy servers can be consolidated on a single platform, consuming less electricity, generating less heat (and demand for a/c) and wasting less resources - under-utilized capacity. IT departments are liberated to focus less on individual system needs and more on overall quality of service. Priority can be given to the most important applications, since resources such as CPU cycles and memory are not tied each virtual server's physical hardware. The virtual servers have no physical hardware - just their share of the total hardware available.

The agility of virtual servers, such as the ability to migrate between physical hosts without stopping, is an enabling technology for greatly improved quality of service, high-availability solutions, performance tuning, and disaster recovery. Administrators can back up entire ready-configured systems rather than only the critical data. Test-beds can be created and decommissioned at will. Critical servers can recover on totally different hardware. Planned maintenance can unfold during normal hours without any interruption of service.

Virtual Applications

One of the most time-consuming IT duties can be the maintenance of workstations throughout the network. With so much software to install and configure, it can take several hours of work to thoroughly configure a new workstation for a new hire. Company-wide application upgrades can be be quite burdensome. There are quite a number of approaches to this.

Most enterprises use a combination of centralization, standardization and centralized management.

Centralization can include hosting of files on a server so all important files are in one place. It can also include centralization of applications, for example, a preference towards web-based applications so that all that is required from the workstation is a functioning browser. Traditionally workstation-centric applications such as word processing can also be centralized, and IT departments are doing so in increasing numbers.

Application centralization offers the benefits of centralized management, centralized maintenance, enterprise-wide consistency, and a consistent user experience, regardless of the user's location or how they have connected to the application server. The importance of workstation performance is greatly reduced, and workstations can be commissioned or replaced very quickly. In highly centralized environments, it is common for a maximum time limit for troubleshooting any workstation issue since any workstation can replaced with a "known good" substitute in as little as fifteen minutes. Why then, would a support department delve deeper into mysterious problems?

Virtual IT and Software as a Service

Since pools of processors and highly-reliable storage can now be freely allocated between servers that need them, it's not surprising that vendors have begun to offer power-on-demand services and applications on demand. Business owners now have a choice between buying IT assets and subscribing to IT services.

A business can acquire virtual servers on a month-to-month basis, effectively delegating much of the day-to-day management of the systems to a company that specializes in managing a server farm, large quantities of data and raw processing power. Backup services (Cloud Storage) are an option for secure off-site storage as part of a disaster-recovery plan. Software as a service (SAS) is a valid alternative to actually purchasing licenses and supporting applications internally. Virtual Servers can be provisioned by third parties for a business`s exclusive use.