Franco R. Negri

Subscribe to Franco R. Negri: eMailAlertsEmail Alerts
Get Franco R. Negri: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Article

Operational Management: A Web Service Barrier-to-Entry

Operational Management: A Web Service Barrier-to-Entry

Web services represent a quantum leap forward towards achieving not only interenterprise application integration, but also cross-enterprise application interoperability. These services will enhance operational efficiency and agility, and unlock the intrinsic value of the company's goods and services, thereby creating new market opportunities.

However, as organizations develop initiatives to electronically expose their core competencies and express them as Web services, they are confronting some specific challenges. The first is that most organizations may not have considered the need for a paradigm shift in their IT departments toward a more e-business service-delivery approach. The second is that current management tools look only at technology- and component-based metrics apart from the concept of quality e-business service delivery. Both of these challenges make it difficult to successfully implement Web services in an organization.

Defining "Service"
To frame the discussion, let's first define "service": a) the occupation or function of serving, b) a facility of supplying some public demand, and c) contribution to the welfare of others. Web services fulfill public demand by using technology to facilitate consumer interactions (consumers can be internal or external to the organization). That is, they act as an abstraction between service consumers and the application and infrastructure components which provision them. Web services synthesize technology into electronic business deliverables - electronic manifestations of business processes. They distill the interactions of service consumers into single points of electronic contact. A Web service can be further defined not only by the value it exposes but by its repeatability and reliability.

As an example, let's suppose there is a Web service that makes it possible for brokerage firms to inquire into a customer's credit history. The value and viability of this Web service could be described not only by the information it exposes, but by the timeliness of the response to the brokerage's request, the accuracy of the information, and whether the service can consistently facilitate the needs of the consumer. Essentially, a Web service that provides all the required information quickly and consistently over time, defines the service itself. Therefore, implementing a Web service requires more than just the technical task of developing and provisioning it - it also requires adhering to standards of consistent performance and reliability. Another way of saying this is that it is as important to implement an operational environment of high-quality, consistent execution around Web services as it is to create the services themselves.

Essentially, Web services represent the dividing line between infrastructure (the domain of IT management) and consumers (the domain of business line management). From this, we can draw two important conclusions. First, Web service quality is the single biggest factor in determining customer satisfaction, and thus the most meaningful target for managing and controlling service levels. And second, from an operational perspective, the necessary alignment of business and IT management can only occur if both are focused on common service delivery objectives.

Shifting IT to a Business Focus Using Service-Level Management
It's clear that Web services and their ensuing deployments will affect not only development staff, but also IT operations and business line managers. Web services will require an institutional shift that involves new service-centric processes, methods, and tools that can be shared effectively across these functional areas to manage and assure service quality.

While most development shops have figured out how to technically leverage Web services, few have considered the impact these initiatives will have on those parts of the organization that are chartered with maintaining them. For Web services to be successful in their ability to provide application service interoperability inside and outside the organization, there must be a concomitant shift in the IT operations domain away from technology-focused operations toward business-focused operations. New software solutions to enable this change are in order.

Traditionally, the coordination of activities between IT development, operations, and lines of business have been difficult and a barrier to the "agility" that everyone speaks about. The proliferation of Web services will only exacerbate this problem as business line organizations request more Web services and demand accountability, while IT organizations continue to be resource constrained.

The good news is that Web services will act as a catalyst for institutional change and inevitably forge a closer bond between IT and business. IT organizations will need to engage in a new form of management, namely, defining and managing services to strict business-centric guidelines. They will need to manage and monitor technology to the objectives of business. In practice, service-level agreements are the vehicle for accomplishing this. They will be used to describe the metrics of quality in accordance with business objectives for the service. They will also act as an enabler to transition IT from technology-centricity to more business-focused operations.

Additionally, as Web services are implemented and become commonplace, IT operations will be tasked with not only managing their own organization's Web services, but also external Web services which their own services rely on. Service-level agreements (SLAs) between organizations will play a vital role. Clearly, this implies that standards and methods for defining SLAs will need to be created. These standards would then be used between IT and business lines internally, and between organizations consuming each other's Web services. Without such service level agreements between departments and organizations for assuring some degree of service quality, it is difficult to imagine a world of interdependent and business-critical Web services.

With that said, SLAs do exist today but are generally technology-oriented. They are usually more focused on network service availability than they are on describing and measuring service quality against business standards. Future SLAs must be defined using both business and technology performance metrics, since both characteristics are required to fully describe the health and viability of Web services.

One way of shifting IT operations to a more service-centric model is to empower them with tools that translate the language of infrastructure performance monitoring into business impact and service-level management. This necessary generation of software solutions will change the way organizations look at application and infrastructure performance by tying it directly to the business services they provision.

Why Traditional Management Tools Don't Work for Web Services
Current management tools and approaches are fundamentally incompatible with the Web services paradigm because traditional tools manage infrastructure and application metrics divorced from the key performance indicators that define service quality. A typical management tool will look at performance metrics for specific tiers of the complex application architectures - database, Web server, application server, or network. They can drill deeply into what might be wrong with a specific component, but can't correlate that problem with the overall health and performance of the Web service that relies on it. This makes it nearly impossible for an IT organization to understand the relationship or impact of technology to service delivery.

Web services-enabled applications will further challenge traditional application management tools. While many Web applications and their associated architectures are n-tiered and inherently dynamic in nature, traditional management tools were designed to manage more monolithic and static computing environments. Traditional management tools don't detect problems, they just report raw data as most approaches use non-intelligent static thresholds as a mechanism for detecting and reporting problems. They also don't attempt to monitor applications and services holistically across all related tiers. These approaches just won't work for managing today's multi-tiered Web applications and services, since it makes no sense to manage a dynamic environment with static thresholds. It is also inappropriate to manage applications component-by-component without regard to how these components work together to deliver services.

The result of not looking across tiers and relying on static thresholds is a noisy and inaccurate management system, which provides numerous false alarms and notifications that might not be relevant to any service context. Thousands of superfluous alarms lead to either an apathetic approach to operations management or to an intensely reactive, finger-pointing, and firefighting approach. Remember what happened the last time you asked someone to explain why an application was running slow?

So a major shortcoming of traditional management tools is their inability to get to the root cause of a problem. Because they focus on monitoring the individual health of components, they are not looking at the health of the entire e-business process. The individual components could look fine, but the overall process could still be broken. This inevitably leads to guesswork and finger-pointing as IT personnel shift blame and avoid responsibility.

The lack of insight into the root cause of problems can lead to over-expenditure on redundant resources just to ensure that the system runs. Without an understanding of where breakdowns and failures occur, redundancy is the insurance policy. Unfortunately, excessive redundancy further contributes to the complexity of the infrastructure and is not usually a good long-term solution.

There are many documented cases where organizations attempted to retrofit legacy management approaches to Web applications and services, and the results were inevitably disappointing, even disastrous. This is solely because traditional management tools and their approaches are based on solving the wrong problems, assuming of course that your ultimate goal is to manage service quality.

The New Generation of Management Tools
To provide meaningful management of Web services, operations staff and business line managers alike require a new breed of evolved management solutions that understand the intricate dependencies and interactions of application components to deliver Web services. This way, management functions are performed in accordance with service-quality objectives instead of component-by-component without regard to service context.

These new service-centric management solutions must monitor the behavior of Web services from the top down instead of the bottom up. They must also incorporate root-cause-analysis capabilities so that Web services and their constituent parts can be monitored as cohesive and federated entities. They must, in essence, manage infrastructure health and performance from the perspective of Web service quality. Anything less and the result will likely be a reactive finger-pointing exercise with excessive operational costs to IT and opportunity costs to business.

New service-centric management solutions will proactively assist operations personnel to answer tough questions in real time. How is a performance anomaly affecting service delivery relative to committed service levels? Which business processes will be affected? These questions cannot be answered without understanding the relationships between infrastructure components and the Web services they are facilitating. This understanding will promote efficient operations and the avoidance of costly, unnecessary infrastructure expenditures.

These new tools will also make it possible to realize the paradigm shift needed in IT operations by enabling IT staff to adopt methods and procedures that are focused on service quality. Instead of measuring network availability, or throughput, without service context, IT will interpret these downstream infrastructure metrics within the context of impact on business service objectives. IT will finally be able to align itself to business using service level agreements as the lingua franca. It will also be able to use this mechanism for managing the performance and viability of externally provisioned Web services.

The true test of Web service-oriented management tools will be their ability to intelligently correlate application and infrastructure health and performance information to Web service performance, in real time. The rate of success, speed, and adoption of Web services will be directly tied to organizations' adoption of tools to effectively manage them.

More Stories By Franco R. Negri

Franco Negri is the founder, CTO, and chief strategist of PANACYA. In a 23-year career with leading suppliers and consumers of advanced management technology, Franco developed a keen understanding of market needs and a strong vision for the next generation. He was most recently VP, Product Marketing and VP, Research & Development at Computer Associates, where he was responsible for Unicenter TNG - CA's flagship Enterprise Systems Management product line.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.