ArticlePDF Available

Abstract and Figures

It is critical to address potential performance problems of XML-based SOAP to design a Web Service communication framework for mobile devices, as the mobile computing environment is physically constrained to limited computing power and network bandwidth. In this paper, we describe a Web Service communication framework for mobile computing, which optimizes SOAP message contents. The performance and efficiency of Web Service messaging can be significantly improved by removing the redundant or unchanging parts of SOAP messages. This paper shows that the redundant or static parts of a SOAP message may be treated as metadata and stored in a shared metadata space, for which we integrate our optimized Web Service communication framework with an information service. We evaluate our architecture through a benchmark testing of the resulting system. The empirical result shows that we save on average 83% of message size and on average 41% of transit time.
Content may be subject to copyright.
Architecture for High-Performance Web Service Communications
Using an Information Service
Sangyoon Oh(1,2)*, Mehmet S. Aktas(1,2)*, Marlon Pierce(1), Geoffrey C. Fox(1,2)
(1) Community Grids Lab, Indiana University, Bloomington, Indiana, 47404, USA
(2) Computer Science Department, School of Informatics, Indiana University
{ ohsangy, maktas, mpierce, gcf }@cs.indiana.edu
* Corresponding Authors
Abstract: - It is critical to address potential performance problems of XML-based SOAP to design a Web Service
communication framework for mobile devices, as the mobile computing environment is physically constrained to
limited computing power and network bandwidth. In this paper, we describe a Web Service communication
framework for mobile computing, which optimizes SOAP message contents. The performance and efficiency of
Web Service messaging can be significantly improved by removing the redundant or unchanging parts of SOAP
messages. This paper shows that the redundant or static parts of a SOAP message may be treated as metadata and
stored in a shared metadata space, for which we integrate our optimized Web Service communication framework
with an information service. We evaluate our architecture through a benchmark testing of the resulting system. The
empirical result shows that we save on average 83% of message size and on average 41% of transit time.
Key-Words: - Grid/Web Service, WS-Context, Mobile Applications, Information Systems, Quality of Service
1 Introduction
The SOAP message enables applications on
heterogeneous platforms interoperate with each other
by defining text-based remote procedure call (RPC)
mechanism. However, the verbose nature of a SOAP
message holds potential overheads. For example,
when data is converted to and from a SOAP message,
both size and processing time of the message is
increased substantially. This creates performance
inefficiencies in some application domains, such as
mobile computing. The mobile computing
environment, which holds many physical constraints
like limitations in processing power, battery life, and
wireless connections, needs an efficient solution to the
problem of expensive processing cost of SOAP
messages. We proposed and implemented a research
framework, which is designed to provide efficient and
optimized message exchange paradigm in mobile Web
Service environment, the Handheld Flexible
Representation (HHFR) [1] [2]. By using the HHFR
architecture, participating applications can a)
exchange messages in flexible presentation, such as
binary format, and b) optimize message contents by
removing the redundant parts of the SOAP message.
In this paper, we discuss the use of a Context-store,
which is a meta-data repository for HHFR, and present
our particular investigations and experiences of using
information management research framework, Fault
Tolerant High Performance Information Service
(FTHPIS) [3] [4], as the Context-store. In this research,
the redundant message parts of the SOAP message,
which is exchanged between participants, are treated
as metadata and stored in the Context-store. This way,
the size of exchanging SOAP messages is being
minimized. The redundant parts of a SOAP message
can be considered as XML fragments, which are
encoded in every SOAP message. These XML
elements are stored as context, i.e. metadata associated
to a conversation, into the Context-store. Each context
is referred with an URI. The uniqueness of the URI is
ensured by the system. Upon receiving the SOAP
message, the corresponding parties may interact with
the WS-Context Specification [5] compliant
Context-store to retrieve the context associated with
the URIs in the SOAP message. In addition to
minimizing message size, the use of the Context-store
guarantees integrity of exchanging messages. For
example, a late-joined participant should access and
retrieve stored negotiated information from the
Context-store to understand an ongoing conversation
between a service and other participants. Our
presentation is particularly focusing on applications in
mobile computing environments, but the approach
may be more general.
We describe here a novel approach to minimize a)
the size of SOAP messages, b) the cost of handling
XML messages for high performance Mobile
Grid/Web service applications. We discuss our
implementation, which uses WS-Context-complaint
Information Services as a Context-store.
This paper is organized as follows. In section 2, we
discuss relevant works. We discuss our architectural
and implementation details in Section 3. In Section 4,
we present the performance evaluations of proposed
approach. We summarize and discuss the future work
in Section 5.
2 Background
In this section we overview previous efforts that focus
on improving Web Service communication
performance. We also discuss existing researches
focusing on standardization of web service
communications.
2.1 The lineage of SOAP alternatives for
mobile Web Services
Because of increasing demand of binary form of
XML-Based communication methods, W3C
Workshop [10] was held and produced the report on
Binary Interchange of XML Information Item Sets
(Infoset) [7]. The report includes conclusions of the
workshop meeting on September 2003 as well as
several dozens of position papers from various
institutes [11] – [13]. The purpose of the workshop
was to study methods to compress XML documents
and transmit pre-parsed and schema specific object. It
identified requirements of binary XML Infoset, for
examples a) maintaining universal interoperability, b)
producing a generalized solution that is not limited to
a specific application domain, c) reducing process
time including a data binding time, and d) negotiation
about falling back to XML/SOAP text format if
receiver can’t understand binary. Web Service
performance has been more recently reviewed at the
15th Global Grids Forum workshop (GGF 15) [22].
We put current approaches of improving Grid/Web
Service communication performance into different
categories. First, most proposals that follows the W3C
XML Binary Characterization have a goal of
producing a self-contained alternative to an XML
message, which is optimized for faster processing and
has smaller packet size. Approaches in this category
replace a redundant vocabulary with indexes. Sun's
Fast Infoset project [14], XML Schema-based
Compression (XSBC) [15], XML Infoset Encoding
(XBIS) are several examples of the category. The
second category is a non self-contained alternative,
such as Sun's Fast Web Services [11] and the Indiana
University Extreme! Lab’s recommendation [16]. Our
HHFR also falls on to this category. The last category
is a message compression approach. Compression
reduces the size of a XML document, but increases
processing time. XML-specific compressions like
XMill [17] achieve better ratio than conventional
compression utilities like GZip [18]. But even XMill
doesn’t improve performance much because of the
additional layer of processing – compression and
decompression.
The Global Grid Forum’s Data Format Description
Language (DFDL) [19] is a descriptive language. It is
proposed to describe a file or a stream in a binary
format for Grid computing. Like the older Extensible
Scientific Interchange Language (XSIL) [20], it is
XML-based and comes with an extensible Java Data
model. DFDL defines the structure of data. For
example, it defines a number format of data, such as
whether it is a big-endian or little-endian, and a
complex data format such as an array. Also DFDL is
designed to be processable through a DFDL parser and
its data model. We designed the message format
description of our Flexible Representation based on
DFDL. In our Handheld Flexible Representation
architecture, we define simple XML-schema based
descriptive language and develop a language parser
using XML Pull Parser (XPP) [21]. Our prototype
implementation is not as in-depth as DFDL, though it
will be enough to show advantages of our approach.
The HHFR architecture is categorized the non
self-contained alternative approach and it focuses on
optimizing message stream that we believe the most
appropriate approach for mobile Web Service
applications with high latency connections and limited
computing power.
2.2 Specifications for Grid/Web Service
interactions
Often an application for a specific purpose is
composed of multiple Grid/Web Services. For
example, an airline reservation system could consist of
several Web Services, which are combined together to
process reservation requests, update customer records,
and send confirmation. To standardize the way of
sharing information or meta-data between multiple
collaborating services, a specification is needed. There
are many specifications, such as Web Service
Composite Application Framework (WS-CAF) [6],
Web Service Resource Framework (WSRF) [7], and
WS-Metadata Exchange [8]. They are introduced to
define stateful interactions among services.
The WS-CAF specification is a collection of three
specifications, WS-Context, WS-Coordination
Framework (WS-CF), and WS-Transaction
Management (WS-TXM). WS-Context specification
defines a simple mechanism to share a common
context for multiple participating Web Services. A
participating application can discover results of other
participants’ execution, which is stored as context.
WS-CF defines a coordinator, which makes a context
operation persistent and message delivery guaranteed.
WS-TXM defines three distinct transaction protocols:
two phase commit, long running actions, and business
process flows. They make existing transaction
managers interoperable. Three specifications
comprise a stack of functionality. WS-Context is at the
bottom and adding WS-CF and then WS-TXM.
Web Service Resource Framework (WSRF)
specification, which is proposed by Globus alliance,
IBM, and HP, is an alternative specification to
WS-CAF [9]. WSRF defines conventions for
managing state, so that collaborating applications
discover, inspect, and interact with stateful resources
in standard and interoperable ways like WS-CAF.
Web Service Metadata Exchange (WS-ME) provides
a mechanism a) to share information about the
capabilities of participating Web Services and b) to
allow querying a WS Endpoint to retrieve metadata
about what to know to interact with them. The
combination of two specifications would achieve what
WS-CAF targets. But the combination of WSRF and
WS-ME have limitations to accomplish state
management, since WSRF just enables access and
update rights and WS-ME defines only how to access
interaction independent metadata.
Among the existing specifications, which
standardize service communications, we choose
WS-Context Specifications to design our architecture.
Unlike the other service communication specifications,
WS-Context models a session metadata repository as
an external entity where more than two services can
easily access/store highly dynamic, shared metadata.
3 Architecture Overviews and
Implementation
The architecture of the proposed approach consists
of two individual systems: HHFR and FTHPIS, as
depicted in Figure 1. The HHFR architecture provides
layers, which optimize and stream messages to
achieve high performance mobile Web Service
communication. HHFR uses FTHPIS as the
Context-store to store redundant or unchanging parts
of messages. It should be noted that HHFR uses only a
part of FTHPIS’s functionalities, WS-Context
compliant functions. These functions enable both
participants of mobile Web Service conversation and
the HHFR System to store unchanging parts of the
communication state, such as the related message
schema and the HHFR design specification schema.
3.1 An Overview of Handheld Flexible
Representation (HHFR)
As we discuss briefly in the introduction, the HHFR is
a software architecture designed to provide an
efficient mobile Web Service communication. There
are three key features of the architecture: a) separation
of message contents, b) streaming messages, and c)
the use of Context-store.
3.1.1 Separation of message contents
Similar to other alternatives of the conventional
SOAP-based Web Service communications, one of the
key issues of HHFR architecture is optimizing
conventional SOAP messages by separating redundant
parts of the message content from SOAP messages.
Some advantages of using optimized SOAP message
format can be summarized as a) reduced bandwidth
consumption, b) reduced message transit time and c)
reduced time required for processing SOAP messages.
By simplifying the structure of messages, our aim is to
remove text parsing and serializing overhead. We
Figure. 1. Overview Of The Architecture
achieve optimized message format/representation by
two steps. First, HHFR separates message contents
from its syntax (i.e. XML type and structure) and then
convert the message content into a preferred
representation formed from the application’s
in-memory representation. To this end, one can avoid
text-to and text-from data conversion, which
consumes lots of process cycles. Second, HHFR
streams message with preferred representation. To
map non-XML based data (i.e., data separated from
SOAP message content), and XML data (i.e., SOAP
message itself), we define a DFDL-style data
descriptive language, termed as the HHFR schema. It
is a small subset of XML Schema Definition (XSD)
[27] with some additions.
3.1.2 Streaming messages
The HHFR approach works best for stream of
messages. For example, messages between a specific
service and a client might have the same structure and
data type, if the client keeps requesting the same type
of information. Although, the values in the message
may change, the structure and type of data will remain
the same during consecutive transactions, which we
call a stream. Therefore the structure and type of
messages in the stream can be transmitted only once,
and the rest of messages in the stream contain only
payloads. The structure and type of the message are
defined in HHFR Schema language. To establish such
message stream, two endpoints should negotiate about
the characteristics of the stream at the beginning of the
session. In this case, the two endpoints negotiate a
preferred representation (a binary representation), a
transport protocol (TCP or UDP), and quality of
service issues (messaging reliability and/or security).
The negotiation uses a conventional SOAP message,
so that two endpoints fall back to conventional SOAP
messaging when the negotiate fails.
3.1.3 The use of Context-store
The Context-store, which is our research focus in this
work, is a repository for storing redundant or
unchanging parts of the messages in the stream and the
context of the stream, such as message structure and
type as a HHFR Schema document and characteristics
of the stream.
3.2 An Overview of Fault Tolerant High
Performance Information Service
(FTHPIS)
Information Services may be thought of as a solution
to general problem of managing metadata about
Grid/Web Services. The benefits of using an
Information Service in Mobile Computing
Environment are two fold.
The first advantage is related to dynamic discovery
capabilities of mobile Web Services. Mobile
computing environment presents fragile
communication characteristics, such as intermittent
wireless connections and low bandwidth. These
characteristics cause high communication failure rates
and limitations in fast message exchanges. Such
limitations require a solution where each mobile
device can interact with an available service which
may satisfy the requirements of fast and reliable
communication between mobile devices and mobile
Grid/Web Services. Also, mobile services have a
volatile behavior. In other words, these services may
come and go. It is important to locate mobile services
in a dynamic fashion. In order to solve these
limitations, an Information Service is required to
dynamically locate the most reliable and high
performance mobile services among those which are
of the same type.
The second advantage is related to providing fast
and efficient Web Service transactions in order to
satisfy fast and reliable communication requirements
of mobile computing Environment.
We have designed and implemented Fault Tolerant
High Performance Information Services (FTHPIS)
that supports both dynamically generated and
semi-static metadata. As depicted in Figure 2, we have
utilized two existing Web Service standards
(Universal Description, Discovery, and Integration
HHFR Client Service Provider
(HHFR capable)
Extended UDDI
Service
Information Service
(FTHPIS)
(2) Service
Discovery Request
(3) Service
Information
(1) Service
Register
(4) HHFR Session
WS-Context
Service
Figure. 2. UDDI Usage For Service Discovery
(UDDI) [23] and WS-Context) in our implementation.
We have extended existing UDDI Specifications and
designed an extension to UDDI Data Structure and
UDDI XML API to be able to associate both
prescriptive and descriptive metadata with service
entries [24]. We used name-value pairs to describe
characteristics of services. Apart from the similar
methodologies [25], [26], we provide both general and
domain-specific query capabilities. An example for
domain-specific query capability could be
XPATH/RDQL queries on the auxiliary and
domain-specific metadata files stored in the UDDI
Registry. As depicted in Figure 3, a possible scenario
where FTHPIS may be used in mobile computing
environment is as follows: 1) a HHFR capable service
provider registers its service to FTHPIS, 2) HHFR
client sends discovery request to locate available
HHFR capable services, 3) FTHPIS response with a
list of services matching the request, 4) HHFR client
initiates the session with one of the available services.
We have also implemented extended version of
WS-Context Specifications to provide an interface for
publishing and accessing session metadata. This way
the FTHPIS supports not only quasi-static, stateless
metadata, but also more extensive metadata
requirements of interacting systems. Here, an
Information Service forms a hybrid XML metadata
store for both dynamic, high updated session data and
static/semi-static, rarely changing metadata.
We use our FTHPIS as a medium between
communicating parties to store redundant parts of the
XML based SOAP messages. Here, the Information
Service keeps track of context information shared
between multiple participants in Grid/Web Service
interactions. It also maintains user profiles and
preferences, application specific metadata,
information regarding sessions and state of entities in
these sessions. A context may contain arbitrary
information which in turn enables sessions to be
shared across multiple services and shared session
content to be changed over time. A context might be
associated to either to a service or the session itself. In
this research, our main focus is session related context
which is generated dynamically during the
conversation of participating entities in a session.
3.3 Implementation of Mobile Web Service
application with Context-store
As the FTHPIS system is implemented as Apache
Axis compatible Web Service, a general client to the
FTHPIS system requires a SOAP-Java binding of the
Apache Axis [28] library. The Axis implementation
for the J2ME environment has not been developed yet
and is not expected to be developed in the near future
because of the lack of related programming libraries,
such as advanced XML parsers, utility libraries, and
memory space limitation. Poring the existing client
interface code in Java 2 platform, Standard Edition
(J2SE) [30] to Java 2 platform, Micro Edition (J2ME)
[29] is not a viable approach because of limited
functionalities of J2ME. Due to these limitations, we
choose an alternative approach which serializes a)
SOAP request messages direct from an in-memory
representation and b) parses response messages with a
simple SOAP parser without the Axis’ Java-SOAP
binding. To do this, we use the kSOAP [31] library,
which is an open source SOAP library for J2ME
applications.
We use the two primary WS-Context related
functions, getContent() and setContent() methods,
to access and store redundant information into the
Context-store. In the implementation, we made
method calls from mobile clients not tied to any other
operations in the HHFR session by making our
implementation as multi-threaded. This way, these
methods may be called at anytime when the HHFR
runtime or the HHFR client service needs to create,
update or retrieve context in the Context-store. The
following Java program a) creates a
ContextServiceHandler object with the Context
Service URl and the version of the service as
parameters, b) stores a given context paired with a
unique identifier, and c) retrieves context.
Context-store
(Information Service)
getContent()
Mobile Context Service Client
ContextServiceHandler
setContent()
kSOAP
Conventional Context Service
Client
ContextServiceHandler Axis
getContent() setContent()
As the reader observe in the code sampled given
above, the getContent() and setContent() methods
throw an interrupt exception
(java.lang.InterruptedException), which is
thrown when a thread is waiting, sleeping, or
otherwise paused for a long time. As the handler runs
as a thread, it avoids a possible deadlock situation,
which may be occur when network failed or
operational error.
The ad-hoc method to generate a SOAP message is
the biggest obstacle to automate client code generation,
compared to automatic Java binding generation of
Axis. The current implementation of using FTHPIS
Service as a Context-store holds a limitation in high
level Java-binding Interface to SOAP message for
HHFR clients due to limited programming library of
mobile computing.
4 Evaluation
In this section, our main goal is to investigate the
performance of proposed approach. In the system
evaluation section, we particularly address the
following questions:
· How much performance gains by using a
Context-store in Web Service Messaging?
· What is the baseline performance of the
Context-store implementation (the FTHPIS
system) from the perspective of a mobile client?
4.1. Evaluation of Performance Measurement
With and Without Context-store Usage
In this experiment, we measure the bandwidth gain
from reducing the size of message. Our choice for a
sample SOAP header is from the WS-Addressing
Specification [32]. The WS-Addressing Specification
defines a transport neutral mechanism to address Web
Services and messages. It defines two constructs,
endpoint references and message information headers,
to convey information between Web Service
endpoints, the target of Web Service messages.
Participating nodes store unchanging SOAP parts to
the Context-store (Information Service) and retrieves
them when it is needed. So we can store most of
WS-Addressing header parts to improve message
communication performance. A practical example of
this usage case is as following: two Web Service
endpoints: A, which is a service provider and B, which
is a mobile Web Service client start a series of Web
Service transactions. Endpoint B requires
WS-Addressing headers only if it needs to send a reply
or addresses an individual message. So the header is
archived in the Context-store. Among elements of
WS-Addressing headers, <messageID> shouldn’t
be archived because it is unique for each message. The
example scenario is depicted in Figure 4.
Since a purpose of the experiment is to show how
much performance gain in time by adopting the
Context-store, we measure the performance of a
HHFR transaction, not a conventional Web Service
transaction. It is because messages are exchanged as a
HHFR stream after a negotiation stage, which includes
a context-saving process at the beginning.
As depicted in Figure 4, a service we use is a simple
echo service for experiment, which immediately
returns a received message back to a sender. First, we
measure round trip times (RTTs) of echo transactions
with a full WS-Addressing header over HHFR
communication channel, which simulates a without
the Context-store case. Then, we measure RTTs of
echo transactions with a minimized header, which
simulates a with the Context-Store case. Since we are
Figure. 4. A scenario for WS-Addressing example
ContextServiceHandler handler =
new ContextServiceHandler(SERVICE_URL, 0);
try {
boolean result =
handler.setContext(identifier, context);
Object contextData =
handler.getContext(identifier);
}
catch (java.lang.InterruptedException exception) {
exception code…
}
interested in a one-way transit time, we divide result
RTTs by two.
The results show we save on average 83% of
message size and on average 41% of transit time by
using our approach. The summary of the result is
shown in Figure 5. The scenario is simplified to
remove possible Context-store accesses during a
stream session. So our experiments are results of the
best case scenario.
MIDP’s System.currentTimeMillis() call is
used for timing measurements. Table 1 contains a
summary of testing environments.
Table 1 Summary of Machine Configurations
Context-Store: GridFarm 8
Processor
Intel® Xeon™ CPU (2.40GHz)
RAM
2GB total
Bandwidth
100Mbps
OS
GNU/Linux (kernel release 2.4.22)
Java Version
Java 2 platform, Standard Edition
(1.5.0-06)
SOAP Engine
AXIS 1.2 (in Tomcat 5.5.8)
Service Client: Treo 600
Processor
ARM (144MHz)
RAM
32MB total, 24MB user available
Bandwidth
14.4Kbps
OS
Palm 5.2.1.H
Java Version
Java 2 platform, Micro Edition
CLDC 1.1 and MIDP 2.0
4.2. Evaluation of Performance Measurement
With and Without Context-store Usage
In this experiment, we investigate baseline
performance of the Context-store (FTHPIS with
version 0.9v1) accessing overhead. We use a Dummy
Web Service as a yard stick to compare results of the
Context-store access. It echoes back a received
message, like the service of the previous experiment.
The purpose of the test is to measure a pure data
delivery-overhead of the given XML fragment (the
unchanging SOAP header) without processing any
Context-store related operation. The experiment is
conducted with the same configuration of the previous
experiment. We also conduct a simple empty message
service that receives an empty SOAP message and
returns it with the same configuration setup. The
scenario of the setContent experiment and two other
yardstick experiments are depicted in Figure 6.
The message used as a sample SOAP header that is
stored in the Context-store is a message example in
Web Service Reliable Message Specification
(WS-RM) [33]. The size of the payload is 847 byte
and the entire SOAP message size is 1.58KB. The size
Figure. 6. The configuration of experiments
(setContent)
Empty (2 byte) Medium (513 byte) Large (2.61 KB)
500
1000
1500
2000
2500
3000
Message Size
Transit Time (msec)
Without Context-store
With Context-store
Figure. 5. Transit time of message exchange: each test
set represents 30 iterations and the average value is
presented on the graph. Messages are exchanged using
HHFR communication channel
of empty SOAP message for the latency test is 359
byte. We measure RTTs of the entire message trip
except a message generation overheads. The
summarized test results are shown in Figure 7.
As depicted in the figure, the Context-store
(FTHPIS) service access adds up minimal or no
overhead to the overall performance by comparing its
results with Dummy Web Service test results, because
the only difference between two experiments is
whether the service accesses the Context-store service.
We note the time difference between two experiments
is less than 100 msec, which we claim the
Context-store access overhead. The major factor that
has a linear relationship with RTTs is the size of the
message. Messages in the Context-store and the
Dummy service test are 1.2KB bigger than the Empty
message test, and this causes big time differences
between two groups.
5 Conclusion
We investigated a novel approach to optimize
Web Service messaging performance, in which
the system stores unchanging or redundant
metadata exchanged in the messages in a
third-party repository, the Context-store. This
approach increases the efficiency of
conversational Web Service message exchange,
since the messages contain many redundant
SOAP message elements in a conversation.
We evaluated the performance gains of
adapting our approach. As anticipated, the
empirical result shows that the message size is
reduced to 17% of original and transit time is
saved in 41% (on average). The Context-store
accessing adds maximum 100msec overhead,
which is 2.5% of adding to the RTT of the same
size message delivery, to RTT of the dummy Web
Service.
Combined with separation of message contents
and message streaming, our approach – metadata
storing to repository (the Context-store)
confirms the feasibility of our experimental
architecture for efficient and fast mobile Web
Service applications.
References:
[1] S. Oh, G. C. Fox, “Optimizing Web Service
Messaging Performance in Mobile Computing,”
Community Grids Lab Technical Paper, Feb.
2006.
[2] S. Oh, H. Bulut, A. Uyar, W. Wu, G. Fox,
“Optimizing communication using the SOAP
Infoset for mobile multimedia collaboration
applications,” in Proceedings of The IEEE 2005
International Symposium on Collaborative
Technologies and Systems (CTS 2005), St. Louis,
Missouri, USA, May. 2005.
[3] Bunting, B., Chapman, M., Hurley, O., Little M,
Mischinkinky, J., Newcomer, E., Webber, J., and
Swenson, K., “Web Services Context
(WS-Context),”
http://www.arjuna.com/library/specs/ws_caf_1-0/
WS-CTX.pdf
[4] M. S. Aktas, G. C. Fox, and M. Pierce, “Managing
Dynamic Metadata as Context,” in Proceedings of
The 2005 Istanbul International Computational
Science and Engineering Conference
(ICCSE2005), Data Istanbul, Turkey, Jun. 2005.
[5] Community Grids Lab, “Extended UDDI and
Fault Tolerant and High Performance Context
Service,” http://www.opengrids.org
[6] E. Newcomer et al., “Web Service Composite
Application Framework (WS-CAF),”
http://www.oasis-open.org/committees/tc_home.p
hp?wg_abbrev=ws-caf
[7] K. Czajkowski et al., “The WS-Resource
Framework”, Globus Project, White Paper, 2004,
http://www.globus.org/wsrf/specs/ws-wsrf.pdf
Set 1 Set 2 Set 3 Set 4 Set 5
0
500
1000
1500
2000
2500
3000
3500
Round Trip Time (msec)
WS-Context Service
Dummy Web Service
Empty SOAP message
Figure. 7. Round trip time of experiments: each test set
represents 30 iterations and the average value is presented on
the graph.
[8] K. Ballinger, et al, “Web Services Metadata
Exchange,” Sep. 2004,
http://msdn.microsoft.com/library/en-us/dnglobsp
ec/html/ws-metadataexchange.pdf.
[9] The Globus Alliance, http://www.globus.org
[10] World Wide Web Consortium, “XML
Information Set”,
http://www.w3.org/TR/xml-infoset/
[11] P. Sandoz, S. Pericas-Geertsen, K. Kawaguchi,
M. Hadley, and E Pelegri-Llopart, “Fast Web
Services”, Aug. 2003,
http://java.sun.com/developer/technicalArticles/W
ebServices/fastWS/
[12] J. H. Gailey, “Sending Files, Attachments, and
SOAP Messages Via Direct Internet Message
Encapsulation”, Dec. 2002,
http://msdn.microsoft.com/msdnmag/issues/02/12
/DIME/default.aspx
[13] D. Brutzman, and A. D. Hudson, “Cross-Format
Schema Protocol (XFSP)”, Sep. 2003,
http://www.movesinstitute.org/openhouse2003slid
es/XFSP.ppt
[14] P. Sandoz, S. Pericas-Geertsen, K. Kawaguchi,
M. Hadley, and E Pelegri-Llopart, “Fast Infoset”,
Jun. 2004,
http://java.sun.com/developer/technicalArticles/x
ml/fastinfoset/.
[15] E. Serin and D. Brutzman, “XML Schema-Based
Compression (XSBC)”,
http://www.movesinstitute.org/xmsf/projects/XSB
C/03Mar_Serin.pdf
[16] M. Govindaraju, A. Slominski, V. Choppella, R.
Bramley, and D. Gannon, “Requirements for and
evaluation of RMI protocols for scientific
computing,” In Proceedings of Supercomputing
2000 (SC2000), Dallas, TX, Nov. 2000.
[17] P. Deutsch, “GZIP file format specification
version 4.3,” May 1996,
http://www.ietf.org/rfc/rfc1952.txt.
[18] H. Liefke and D. Suciu, “XMill: an efficient
compressor for XML data,” In Proceedings of the
2000 ACM SIGMOD International Conference on
Management of Data (SIGMOD2000), pp.
153-164, 2000.
[19] M. Beckerle, and M. Westhead, “GGF DFDL
Primer”,
http://www.gridforum.org/Meetings/GGF11/Docu
ments/DFDL_Primer_v2.pdf
[20] R. Williams, “XSIL: Java/XML for Scientific
Data”, Jun. 2000,
http://www.cacr.caltech.edu/projects/xsil/xsil_spe
c.pdf
[21] Java Community Press, “Java Specification
Requests 173: Streaming API for XML”,
http://www.jcp.org/en/jsr/detail?id=173
[22] Global Grid Forum 15 Community
Activity, ”Web Services Performance: Issues and
Research”
http://www.gridforum.org/GGF15/ggf_events_sch
edule_WSPerform.htm
[23] Mehmet S. Aktas, Galip Aydin, Geoffrey C. Fox,
Harshawardhan Gadgil, Marlon Pierce, Ahmet
Sayar, “Information Services for Grid/Web
Service Oriented Architecture (SOA) Based
Geospatial Applications”, Technical Report,
Community Grids Lab, June, 2005
[24] T. Bellwood, L. Clement, and C. Riegen, “UDDI
Version 3.0.1: UDDI Specification,” Technical
Committee Specification.
http://uddi.org/pubs/uddi-v3.0.1-20031014.htm.
[25] A. ShaikhAli, O. Rana, R. Al-Ali and D. W.
Walker, “UDDIe: An Extended Registry for Web
Services,” In Proceedings of the Service Oriented
Computing: Models, Architectures and
Applications, Orlando, Florida, January 2003.
[26] V. Dialani, “UDDI-M Version 1.0 API
Specification,” Report University of Southhampton,
UK 2002
[27] World Wide Web Consortium, “XML Schema
Definition (XSD),”
http://www.w3.org/XML/Schema
[28] Apache, AXIS, http://ws.apache.org/axis/
[29] Java 2 platform, Micro Edition (J2ME),
http://java.sun.com/j2me/
[30] Java 2 platform, Standard Edition (J2SE),
http://java.sun.com/j2se/
[31] kSOAP, http://kobjects.org/
[32] D. Box et al., “Web Service Addressing
(WS-Addressing), Aug. 2004,
http://www.w3.org/Submission/ws-addressing/
[33] R. Bilorusets et al., “Web Services Reliable
Messaging Protocol (WS-ReliableMessaging),
Feb. 2005,
ftp://www6.software.ibm.com/software/developer
/library/ws-reliablemessaging200502.pdf
Article
Service-oriented communication is a new trend in the industry to enable communication through a service-oriented architecture (SOA) and thereby package communications as services, which opens a new paradigm of Web service-based Voice Over Internet Protocol communication, which is extensible and can be easily integrated in end-to-end SOA solutions. In this paper, we proposed the communication components-based open multimedia conferencing Web service over IP networks and mainly focus on the session and media signalling for multimedia conferencing Web service. Details including the multimedia conferencing multiple parties call control state machines and session management, the state life cycle management, and the signal flows for multimedia conferencing scenarios, also including the Simple Object Access Protocol messages mapping to the asynchronous events. The session and media signalling management also provides the Web services interfaces for triggering conferencing session establishment and is further responsible for all necessary media signalling towards a proper media server. In addition, a prototype has been implemented, and its performance has been measured and analysed. Finally, we give the conclusions and future work. Copyright © 2011 John Wiley & Sons, Ltd.
Article
Full-text available
Geographical Information Systems (GIS) presents data-intensive environment for acquiring, processing and sharing geo-data among interested parties. In order to serve geographical information to users in such environment, Service Oriented Architecture (SOA) principles have gained great importance. In SOA-based systems, Information Services support the discovery and handling of these geospatial services. Some options for Information Services in SOA-based GIS systems include a) the Open GIS Consortium (OGC) Web Registry Service (WRS) and b) the Universal Description, Discovery, and Integration (UDDI). WRS is an OGC standard to discover/publish service information of geospatial services. It presents a domain-specific registry capability for geospatial information. UDDI is domain-independent standardized method for publishing/discovering information about Web Services. As it is WS-Interoperability (WS-I) compatible, UDDI has the advantage being interoperable with most existing Grid/Web Service standards. This study presents an approach combining domain-specific registry capabilities of WRS and WS-I compatible UDDI Specifications. We extend UDDI Information Model to support geospatial services. Our approach supports not only quasi-static, stateless metadata, but also more extensive metadata requirements of rich interacting systems. The implementation of our approach is being used to support a GIS workflow system which is a part of NASA Solid Earth Virtual Observatory (SERVO) Grid project.
Conference Paper
Full-text available
Distributed software component architectures provide promising approach to the problem of building large scale, scientific Grid applications [18]. Communication in these component architectures is based on Remote Method Invocation (RMI) protocols that allow one software component to invoke the functionality of another. Examples include Java remote method invocation (Java RMI)[25] and the new Simple Object Access Protocol (SOAP) [15]. SOAP has the advantage that many programming languages and component frameworks can support it. This paper describes experiments showing that SOAP by itself is not efficient enough for large scale scientific applications. However, when it is embedded in multi-protocol RMI framework, SOAP can be effectively used as a universal control protocol, that can be swapped out by faster, more special purpose protocols when large data transfer speeds are needed.
Conference Paper
Full-text available
The Universal Description, Discovery and Integration (UDDI) is a specification for distributed Web-based information registries for Web Services. UDDI allows HTTP-enabled business services to be published, and subsequently searched, based on their interface. UDDI consists of three components: "white pages" to hold basic contact information and identifiers for a company, "yellow pages" to enable companies to be listed based on their industry categories (using standard taxonomies), and "green pages" to record interface details of how a Web service is to be invoked. UDDI is however limited in scope - allowing white, yellow or green pages to be searched based on a few attributes, and does not provide an automatic mechanism for updating the registry as services (and service providers) change. We implement UDDIe -an extension to UDDI, which supports the notion of "blue pages", to record user defined properties associated with a service - and to enable discovery of services based on these. UDDIe enables a registry to be more dynamic, by allowing services to hold a lease - a time period describing how long a service description should remain in the registry. UDDIe can co-exist with existing UDDI - and has been implemented as open-source software.
  • K Ballinger
K. Ballinger, et al, "Web Services Metadata Exchange," Sep. 2004, http://msdn.microsoft.com/library/en-us/dnglobsp ec/html/ws-metadataexchange.pdf.
World Wide Web Consortium
  • The Globus Alliance
The Globus Alliance, http://www.globus.org [10] World Wide Web Consortium, "XML Information Set", http://www.w3.org/TR/xml-infoset/
  • P Sandoz
  • S Pericas-Geertsen
  • K Kawaguchi
  • M Hadley
  • E Pelegri-Llopart
P. Sandoz, S. Pericas-Geertsen, K. Kawaguchi, M. Hadley, and E Pelegri-Llopart, "Fast Web Services", Aug. 2003, http://java.sun.com/developer/technicalArticles/W ebServices/fastWS/
Sending Files, Attachments, and SOAP Messages Via Direct Internet Message Encapsulation
  • J H Gailey
J. H. Gailey, "Sending Files, Attachments, and SOAP Messages Via Direct Internet Message Encapsulation", Dec. 2002, http://msdn.microsoft.com/msdnmag/issues/02/12 /DIME/default.aspx
Cross-Format Schema Protocol (XFSP)
  • D Brutzman
  • A D Hudson
D. Brutzman, and A. D. Hudson, "Cross-Format Schema Protocol (XFSP)", Sep. 2003, http://www.movesinstitute.org/openhouse2003slid es/XFSP.ppt