Chapter 16: Practical Considerations

Web services are being implemented by many enterprises at an intense pace, with little regard for infrastructure concerns such as systems management, high availability, or even interoperability. Many organizations will believe the hype and think they can become application service providers by simply publishing their service in a public registry. While we believe that success in this regard requires more than just technology (a good business plan is what is really needed), we can only provide you with sound, industry-proven technology recommendations.

We have striven to provide you with a complete, 360-degree view of Web services in a manner that will help you architect the right solution for your problem. In prior chapters, the focus was primarily on technology. This chapter will jump into topics, namely systems management, interoperability, payment models, XMLPay specification, service level agreements, testing Web services, performance, availability, scalability, clustering, fault tolerance, and grid computing, that did not fit neatly into other chapters. Its primary focus will be to provide you with best practices and additional considerations for making your architecture become scalable, available, reliable, and on budget.

Systems Management

Many aspects of a comprehensive Web services solution will have components that reside outside the control of IT. Many services, such as credit card authorization, external caching, and public key infrastructure (which is used for authentication and authorization), are vital to the Web services nervous system. The typical Web service infrastructure is composed of a mix of both hardware and software technologies, along with a variety of complex processes to maintain the system's health.

A good systems management strategy allows a centralized, automated approach, to minimize operation errors and ensure compliance with infrastructure policies. This requires both proactive and reactive actions. Proactive monitoring takes into account trends, potential capacity problems, and opportunities for optimization. Reactive actions usually involve monitoring the components in the Web service infrastructure for faults.

Systems management is one solution that can address three business problems: availability, productivity, and efficiency. Availability of a Web service can be defined as the time the service is fully available to internal and external users. Increasing availability may have several benefits, including the ability for internal users to work additional hours and less revenue lost to downtime.

Efficiency of the systems management function can decrease time and travel of support services. This also has the side effect of increasing the user-to-service management ratio. In many organizations today, IT personnel costs and their growth have far outstripped the rate of revenue growth. The time saved in not having to perform user administration, operation functions, and support will free up both management and the technical staff for more proactive tasks.

Systems management requires sound management policies and procedures. Incorporating tight change control to reduce the chance of unexpected results is crucial. A well-thought-out authorization policy, restricting IT permissions to what is needed for personnel to accomplish tasks for which they are responsible, is also a good idea.

Components of a Systems Management Strategy

A complete systems management strategy requires multiple components. Table 16.1 lists management components we feel are required for a Web services environment. Many systems management vendors, such as BMC, CA Unicenter, and Tivoli, may offer additional functionality.

Table 16.1: Components of a Systems Management Strategy

Component	Description
Systems	Enables monitoring of operating systems, disk drives, CPU, and network appliances within a business context.
Performance	Enables monitoring end-to-end performance of a Web service, including bandwidth, network response times, and service performance from the perspective of the end user.
Network	Enables monitoring of network components, such as TCP/IP, routers, switches, hubs, intrusion detection, firewalls, and appliances.
Service level	Enables monitoring of the effectiveness of service delivery and objectives in a manner understandable by the business community.
Scheduling	Enables centralized, event-driven scheduling and error recovery across an organization.
Console	Enables remote control and access from a central location to the components of a Web service infrastructure and eliminates the need to have local staff for routine systems maintenance.
Delivery	Enables the ability to install, configure, remove, and manage software on target servers from a central server. This can include servers, PDAs, laptops, and so on.
Service	Enables the collection and recording of system problems and their resolutions in a centralized database.
Capacity	Enables an organization to make sure that the Web services infrastructure is running at acceptable levels and allows what-if scenarios to help plan for the future.
Backup and recovery	Enables the ability to centrally back up and recover both local and remote servers.