By Michael Bilancieri published: Thursday, August 31 2006
Introduction
Implementing high availability solutions is not a recent trend, especially within larger enterprises. It is, however, becoming a much higher priority for small and medium sized businesses that are now just as concerned with data and application protection as large enterprises.
With today’s more stringent compliance regulations, businesses of all sizes are required to implement security measures to ensure their systems and data are readily available and quickly recoverable. This, combined with the extreme competitiveness of today’s business environment creates a critical need for companies to protect their systems from downtime. If a person is unable to place an order with a company because their system is down, they will simply go to another Website and order it from there. Downtime results in lost business, lost revenues and dissatisfied customers.
Taking a Look Back
For the past 10 years, companies have had the option of implementing a shared-disk clustering solution to help minimize the risk of downtime to their Windows® servers. This alternative is a model where one server monitors another active server and will take over the application should the active server fail. A number of design factors, however, have limited the acceptance and adoption of clustering solutions.
Clusters are complex and require specialized expertise to deploy and manage. This cost can be quite high and restrictive for potential customers.
Clusters leave the data unprotected due to a shared-disk model that introduces a single-point-of-failure. Protecting the data is left up to other solutions such a SAN to ensure redundancy at the storage level, another complex and costly variable.
Cluster solutions require applications to be designed specifically to run within a clustered environment and are not a viable option for many customers because the majority of Windows applications are not ‘cluster aware.’
Smaller businesses, with limited budgets and staff, are typically turned off by the stringent requirements of clusters, leaving them with few options for protecting their applications and their business.
Even after 10 years, Windows clusters still employ the same design with the same inherent limitations and complexities as when they were first introduced. There have been a number of other ‘high availability’ solutions to come along over this time, but most, if not all, use the same limiting, cluster-like architecture.
Beyond Clusters
Today there is another option which utilizes virtualization technology to overcome clustering flaws. This solution enables companies of any size to implement the protection schemes they require. Virtualization isn’t new, having been in use for many years in mainframe systems, storage, networking, etc. It has received a lot of attention recently with the proliferation of server virtualization technologies within the open systems space. Server virtualization is changing how datacenters buy and deploy systems; saving money and improving ROI on server investments. This implementation is using virtualization for consolidation and is not necessarily a true high availability solution.
Another implementation of virtualization, designed explicitly to protect systems and applications from downtime, delivers an availability solution that is far beyond the tired active/passive cluster failover model. This implementation uses virtualization for availability.
Leveraging virtualization provides businesses with a full spectrum of availability solutions; from High Availability for those applications that need to rapidly recover from failures, to complete Fault Tolerance for those critical applications that must run continuously without interruption. These solutions are simple, automated, and affordable, running on standard servers to leverage the preferred choice of hardware.
Using virtualization for the purpose of availability provides many benefits over traditional clusters. These advancements come from the virtualization of two physical servers into a single, fully redundant Windows operating environment. This virtual Windows environment leverages the redundancy of two physical servers to maintain application state even through device and hardware failures. The virtualization layer controls the hardware, error detection, and fault management, masking the hardware so that it is completely transparent to the operating environment and the applications that may be installed. This transparency allows the operating system and any application to be installed without modification. This means no specialized ‘cluster aware’ versions or changes, no scripting, and no failover policies; a complete turn-key solution that supports all applications out-of-the-box.
Benefits of Virtualization
To fully understand the value of this unique approach, you need to think beyond what you know about clusters. Gone are the days of installing, configuring, and managing two fully burdened application environments. No longer is there a need to configure failure management, failover controls, and restart policies. Plus, you can forget about dealing with manual recovery procedures and fail back processes that incur additional downtime. Using a virtual availability solution is akin to managing a standalone reference system, but with all the benefits of completely redundant data and hardware.
When comparing to clustering solutions, the advantages of virtualization are quite substantial:
Single operating environment to manage
Single application license required
Complete protection for the server, hardware, data, and network
No single point of failure – complete redundancy of the storage and data
Can utilize any type of storage – SAN, DAS, NAS
Manageable with basic system management skills
Supports all applications without any modifications
Simple
Less expensive to manage – Reduced TCO
Virtual availability solutions are revolutionizing the availability market. Not only do they offer better choices for the enterprise customer, but with their lower cost and simplified approach, they are answering the demands of the small and medium businesses to which clusters just aren’t an option.
Application availability is no longer just a requirement for the enterprise. The current world and corporate landscapes dictate that all businesses must provide a high degree of application protection and recoverability. Virtualization has come to the forefront of system and application protection, moving beyond traditional clustering technologies and all their limitations. Businesses of any size can immediately realize the benefits of deploying simple and advanced availability technologies within their corporate infrastructures.
With their flexibility, simplicity and reliability, virtual availability solutions are protecting some of the most mission critical systems and applications that we all utilize and benefit from every day. From manufacturing and power plants that require continuous system uptime, to the security systems that control the entrances and exits at our local malls, airports, and office buildings, to infrastructure applications that we all rely on every day including email and databases, virtualization is setting a new standard for availability.
In a Windows availability market that has been stagnant for many years, virtual availability solutions have sparked the dawn of a new era - simple, automated application availability.
Michael Bilancieri is the Director of Products at Marathon Technologies Corporation, the leading provider of high availability software for Microsoft® Windows® environments. For more information on Marathon visit www.marathontechnologies.com.
