Capacity planning addresses
the unpredictable workload of
an e-business to produce a
competitive and cost-effective
architecture and system.
Virgílio A.F. Almeida and
Daniel A. Menascé
An Essential Tool for
Managing Web Services
peed, around-the-clock availability, and
security are the most common indicators
of quality of service on the Internet.Management faces a twofold challenge.
On the one hand, it must meet customer expectations in terms of quality of service. On the other
hand, companies have to control IT costs to stay
competitive.Therefore, capacity, reliability, availability, scalability, and security are key issues to
Web services managers. E-business sites are complex system architectures with multiple interconnected layers composedof many software and
hardware components, such as networks, caching
proxies, routers, load balancers, high-speed links,
and large-database mainframes. The e-business
workload—composed of transactions and
requests submitted to e-business services—is also
complex because of its bursty and highly skewed
load characteristics. Security and authentication
requirements, payment protocols, andthe unpredictable characteristics of Internet service
requests add to the complexity.
For example, it is common for Web sites to
experience, without warning, a manifold increase
in traffic volume. This type of load spike, also
known as a ﬂash crowd, creates terrible performance problems and slow download
times. Such Web delays frustrate customers and cost online business over
$4 billion eachyear, according to a
report from IntelliQuest, a market
research ﬁrm (http://www.intelliquest.
1520-9202/02/$17.00 © 2002 IEEE
That’s why planning e-business service capacity
requires more than just adding extra hardware
based on intuition, ad hoc procedures, or rules of
thumb. Many possible alternative architectures
can implement a Web service; you mustdetermine
the most cost-effective architecture and system.
This is where a quantitative approach and capacity planning techniques come into play. Capacity
planning offers much more than just performance
prediction—it is actually a powerful technique for
managing Web services.
MANAGING WEB SERVICES
The term Web service describes speciﬁc business
functionality exposed by a company, usuallythrough an Internet connection, for the purpose of
providing a way for another company or software
program to use the service.As the Web evolves into
a network of service providers, companies offer
Web-based services to potentially tens of millions of
users via hundreds of thousands of servers. Users
and customers count on the ability to access any
service at any time. Customers’ increasingreliance
on information-based services imposes three
requirements—availability,scalability,and cost efﬁciency—on the services provided by online businesses.
Availability means that users and customers can
count on accessing any Web service from anywhere, anytime, regardless of the Web site or network loads. Availability also means that the site
provides services meeting some measure ofquality, such as a short and predictable response time.
July ❘ August 2002 IT Pro
Figure 1. Capacity-planning process.
Business models and
7. Predict service
6. Calibrate and
2. Characterizethe workload
4. Forecast workload
Scalability means that Web service providers should be
able to serve a fast-growing and unknown number of customers with minimal performance degradation. Cost effectiveness means that the quality of Web services,
represented by availability and fast response times, should
come with adequate expenditures...
Ler documento completo
Por favor, assinar para o acesso.