On the dawn of service-oriented computing, finding relevant Web
services was mainly done by scanning through services registries
(i.e. UDDI Business Registries or UBRs). Automated Web service
search engines were not necessary when Web services were
counted by the hundreds. However, the number of service
registries is gradually increasing and Web service access points
(i.e. WSDLs) are no longer a scarce resource as there are
thousands of Web services disseminated throughout the Web.
Our experiments show building a crawler and a centralized
repository for Web services is inevitable. In this work, we have
used our Web Service Crawler Engine (WSCE), a crawler that is
capable of capturing service information from various accessible
resources over the Web, to help us in conducting our investigation
of Web services on the Web.
In our study, we investigated the distribution of certain elements
and characteristics of the available Web services on the Web.
Distribution based on valid WSDL interfaces, file sizes, HTTP
status, and technology trends are found. Results provide an
overall view on the current status of Web services. An intriguing
result is that fact those search engines have become a new major
source for searching for Web services and that they constitute
72% of Web services available on the Web. Such service statistics
may likely drive search engines to examine the potential of
interoperability with service registries or apply features that can
turn them into effective tools used for discovering services on the
Web.
Although UDDI and search engines provide two distinctive
approaches for finding Web services, it is unclear whether they
will likely merge or coexist. Based on our findings, search
engines have become a new major source for searching for Web
services. Yet, they are vulnerable to returning irrelevant results
and only provide access points to WSDL documents while UDDI
business registries provide a more business-centric model that can