Building in High-Availability: Technology and Services
As organizations move from an eight-hour business day to a 24x365 environment, they need to know the pitfalls of round-the-clock operations. The major influences on application availability are the quality of the application itself and the operational best practices used to manage it.
The cost of downtime and the promise of electronic business contribute to the growing demand for highly available systems. These systems need to deliver computing services to clients on demand at least 99.9 percent of the time during a specified period of time: 10 hours a day, five days a week, for example, or 24 hours a day, 365 days a year. The time period depends on the specific organization.
As organizations move from an eight-hour business day to a 24x365 environment, they need to know the pitfalls of round-the-clock operations. According to GartnerGroup, 40 percent of unplanned downtime is caused by application failure, and another 40 percent is caused by operations. The remaining 20 percent is made up of computer hardware, operating systems, network failures and disasters. According to GartnerGroup, the major influences on application availability – the measure of how often an application is available to deliver an expected service – are the quality of the application itself and the operational best practices used to manage it.
While computer manufacturers and software vendors specializing in operating systems and middleware are using technology to achieve high availability, these efforts only address 20 percent of the problem. Much more work is required to manage the other 80 percent – the application and operational errors. According to the Giga Information Group, "Enterprise hardware continues to get bigger and faster, but hardware alone does not run the system. Enterprise support and services are quickly becoming the key differentiators for companies that require 24x7 operations."
And well they should. The cost of downtime is very high. It’s more than lost revenue – lost information, lost productivity, loss of customer satisfaction and corporate reputation, and disruption of partner relationships can be devastating.
Availability is a function of the total IT environment. The Unisys high availability model represents the IT environment in nine layers, each of which must be specifically addressed if the overall system is to deliver high availability to the user community.
Briefly, the model is composed of:
• Operations, including activities associated with running the data center and the network
• Physical environment, including such considerations as heating, air conditioning, backup power supplies, etc.
• Applications, which are the solutions executing in the environment
• Data, the infrastructure used or created by the applications
• System, including single and multiple software environments
• Hardware, including the capabilities engineered into hardware
• Security, including access issues surrounding highly available systems
• Network, including the communications that deliver response to the end user
Availability can be measured at each of these layers; combined, they comprise the IT environment and determine system availability to the user community. To address each of these layers, Unisys has developed high availability services that can improve the availability of a clients business-critical applications and data, regardless of software, hardware or environmental failures.
High availability services include:
• Environmental readiness reviews to review building facilities, power and air conditioning
• High availability requirements definition of business-critical functions, consequences of disruption, and operational requirements
• Architectural assessments to look at the physical and logical architectures in place and to propose changes that will enable the architecture to meet the high availability requirements
• Clustering to automate application switchover from a failed system to the backup system
• System management reviews of backup and recovery procedures
• Disaster recovery planning to ensure essential data and programs can run at a remote site
• Network performance analysis to identify the networks resilience to failure along with plans to improve the networks overall performance.
• Security assessments to determine the security risks, threats and vulnerabilities of the organization
• Architectural reviews for high-availability storage requirements
Organizations can profit from high availability services at any stage of system development, but they are particularly useful at the outset. Consider a mid-sized bank that is planning to offer round-the-clock electronic banking to its customers via the Internet. To open its accounting systems to customers on a 7x24 basis, the bank must first answer important tactical questions: Can the current IT environment handle the new system with acceptable levels of availability? If so, what services are needed to ensure successful implementation? If not, what new server technology would best support the new application, and how is it best to install, partition and configure the new system? What is the best way to secure the new application from unauthorized access? What degree of security is appropriate? Is an entirely new disaster recovery plan needed, or can the one in place be modified to accommodate the new service?
A High Availability Architectural Assessment answers these questions and others, from system platform configuration to domain design, from network performance to data vulnerability to physical environment considerations. Assessments can include updated or revised Windows 2000 and Windows NT system configurations, recommendations about policies and procedures, and evaluation of training requirements. High availability services can start new business projects off on the right foot and bring them into production successfully.
In addition to individual high availability services, comprehensive new high availability solutions now come with pre-configured hardware, software, and services. Putting all the pieces together in a pre-tested and certified environment allows Unisys to verify system availability and performance. It also makes it possible to offer service level guarantees. Two solutions for which Unisys now offers such guarantees are PeopleSoft and SAP.
Every highly available solution profits from underlying systems that have been specifically designed to offer superior availability and reliability. Until recently, mainframe-class computers were the truly reliable solution platforms, and the availability attributes of "open" platforms in enterprise-wide settings were suspect. Today, however, the Unisys ES7000 Enterprise Server offers mainframe-class availability characteristics on an open server platform. The server is built with Intel processors and Cellular Multiprocessing architecture.
As organizations move to electronic business solutions, continuous system availability is becoming increasingly important. Building high-availability solutions for mission-critical systems, while complex, can be made much easier by using enterprise servers designed for continuous operation, and by securing expert services that both assess and address your overall availability requirements.