The Demands of Speed and Dynamism


The unrelenting pressures of speed and dynamism (constant change) will continue to challenge any organization that depends on the Internet for part of its success and survival. There will always be an advantage to being quick and responsive.

Speed and dynamism are related, but they are also antithetical to each other. Managing service quality would be simpler if things just got faster over time. The traditional Systems Network Architecture (SNA) networks offered by IBM are a good example. The SNA environment remained relatively constant for years. There was growth in traffic volume and the number of connected terminals, but the basic topology and traffic flows remained fixed. Administrators were able to use this consistency to iterate their performance strategies. They could then refine the strategies and apply them to deliver stable, guaranteed performance.

Dynamism makes the challenge more difficult because constant changes present a moving target for planners and administrators. They do not usually have the luxury of working with situations that have long-term stability. Instead, they are forced into making frequent tactical decisions, often working from a changed environment and a new set of trade-offs in every situation.

The emergence of more dynamic service chains, where downstream partners are selected in real time, will force management systems to follow suit. The management systems must dynamically identify each other and establish the rules for communicating. These management system relationships might last indefinitely or endure for time periods as brief as the duration of a single flow. For example, a downstream shipping service might be dynamically selected based on geographic coverage, delivery schedules, and cost. The management system needs to communicate with the selected shipper's management system to detect and help diagnose any problems with transactions involving the shipping service.

The emergence of routing optimization products and services offers multi-homed sites another example of a dynamic service chain. Using these route control products, customers can select Internet service providers (ISPs) in real time, measuring their performance and comparing costs. The basic interactions addressing service disruptions will be supplemented with additional requests for information or new measurements. Other interchanges will focus on reporting trends that indicate a drift toward SLA noncompliance. Efficient providers are rewarded with more traffic and more revenues.

Real-time, customer-to-provider interactions will become more important because customers are always eager to trim their costs and deliver their services more effectively. These factors will continue to push customers and their providers toward real-time exchanges that give the customers increasing control of the resources they buy from the provider.

The term customer network management has been used to describe systems that enable customers to make real-time requests for a range of network services. With customer network management, customers can submit trouble tickets, track their status, obtain billing and usage information, and change service metrics. The real-time interaction can speed problem resolution and enable customers to adjust their bandwidth consumption to their current and projected demands. Service providers also benefit because customers are taking over tasks that were previously handled by provider staff. Service providers can leverage more specialized offerings to strengthen their competitive differentiation.

One obstacle to greater customer participation in services delivery is a distinct lack of integration among the various back-office systems. (These systems include provisioning, billing, order tracking, and capacity planning.) Many new service rollouts, such as digital subscriber line (DSL), overwhelmed the providers and their systems, resulting in long delays and many installation and activation errors. Methods for customer participation in service allocation, accounting, and management exist today in the switched networks used by traditional telephony service providers; as the economics filter to routed networks, the tools and infrastructure will have to catch up.

Customer management systems interact with management systems from business partners, providers, customers, and suppliers. These interactions are between management systems and are different than the business application interactions that are usually defined as business-to-business flows. Different management systems need to exchange information during regular operations, and especially when service disruptions occur.

Because of the complex, dynamic, and rapidly changing web of interactions, adaptable management products are needed. These are tools that discover changes in the managed environment, update their information, and continue operating without staff involvement. The pressures of intense competition, constant change, and growing criticality have outstripped the capacity for hands-on maintenance of service management tools.

If management tools depend on manual entry of information or rules in response to changes in the current environment, they won't be able to keep pace. Organizations cannot afford time delays while they update their tools. They also cannot afford incorrect analyses caused by outdated tools or problems caused by tools failing to provide information in a timely manner.

A simple change, such as assigning an application to a different server, might take no more than a minute, but it can consume hours of staff time to update the management tools if they must be updated manually. The combination of complexity, stringent time pressures, and stiffer penalties for noncompliance is forcing management systems to become more automated with more sophisticated analysis and responses.

RiverSoft introduced its Network Management Operating System (NMOS) to address this problem. (RiverSoft is now part of Micromuse, and NMOS has been integrated into the Netcool product line.) NMOS periodically checks the network infrastructure for changes. When changes are detected, it updates its information accordingly. It then adjusts the correlation information to reflect new connectivity and new dependencies. Other products are taking an intermediate step of detecting and reporting changes, leaving the adjustments to the staff. This feature will be a key differentiator, especially for those organizations that are struggling to stay on top of their own environments.

A similar problem for administrators is the need to understand the relationships between service flows and underlying infrastructures, which is difficult if those relationships keep changing. Without knowledge of the relationships, administrators must take corrective actions without understanding their impacts on the business. They may restore a device, adjust a route, or change access parameters, among other tasks, without knowing if their actions made any difference in the organization's business outcomes. Service management tools must be able to present business-related information to administrators before they make such management decisions.

Mapping these relationships automatically is not an easy task, and the relationships between service flows and underlying infrastructures must be shown in both directions: from flows to elements and from elements to flows. Most management vendors offer a partial solution supporting automatic discovery of elements and applications, leaving the identification of relationships to the staff. This helps, but falls short of what is actually needed.

Dynamism makes the problem even more difficult because continuous changes increase the risk of using outdated information. Some companies I have interviewed are introducing new services weekly, constantly reallocating servers to applications as loads shift, and altering bandwidth assignments in (near) real-time. Perhaps a new approach will be needed that enables applications to register themselves when they are activated. A management tool could receive the registration and update the information accordingly.




Practical Service Level Management. Delivering High-Quality Web-Based Services
Practical Service Level Management: Delivering High-Quality Web-Based Services
ISBN: 158705079X
EAN: 2147483647
Year: 2003
Pages: 128

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net