Reviews:
Sensitivity of Parallel Applications to Large Differences in Bandwidth and Latency in Two-Layer Interconnects by David Vaglia
1.TARGET AUDIENCE
This article is intended for the designers of algorithms and applications used by computers in a parallel architecture. This article is rather high level and focused and considerable background in the areas of networking and distributed is recommended.
2. TYPE OF DOCUMENT
This document is organized as a research project. It includes abstract, background, research, test results, and conclusions.
3. SUMMARY OF ARTICLE
The authors have noted that considerable work and studies have been done on the effects of bandwidth and latency on the performance of applications in a distributed parallel environment. But these systems, for the most part have been relatively uniform as far as the difference in bandwidth and latency in the interconnection between the computers. The authors have asked the question how large differences in bandwidth and latency would affect the performance of the applications. The article describes a test setup involving
several groups of computers organized in a parallel manner. Six commonly known parallel applications were run on the systems and bandwidth and latency were varied to observe the effect on performance. The authors hope that the results from this will help in the development of distributed
applications.
4. CONCLUSION
The most important idea learned from this article is that certain applications and algorithms are more tolerant than others of differences in bandwidth and latency. The application or
algorithm must be studied and evaluated to identify what components cause performance problems. I would recommend this for someone who is designing software to be used in a distributed and parallel environment.
Review #2
http://www.dmst.aueb.gr/dds/pubs/jrnl/1999-MedInf-Euromed/html/euromed.html by David Vaglia
1.TARGET AUDIENCE
This article is intended for designers of modern distributed medical applications and information systems. It is especially intended for European use rather than the United States.
2. TYPE OF DOCUMENT
This article is a proposal for the use of certain types of security standards and and services.
3. SUMMARY OF ARTICLE
This article proposes the use of a Trusted Third Party (TTP) security solution for the concerns of protecting data in medical distributed applications. The article lists a number of security issues and user requirements. Next, a general security model using TTP solutions is described. The article addresses multiple TTP functions. A sample use case is presented. This use case shows an example of a Physician requesting medical data from a remote hospital and the steps involved with registration and security. The paper also touches on more subjects such as legal and financial considerations. The article explains that U.S. cryptographic technology can not be used since it can not be exported from the U.S. This proposal provides an alternative solution for European applications.
4. CONCLUSION
The most important thing in this article is that it shows how many other issues and topics must be considered while addressing security issues. Security technology can potentially affect many current procedures and the designer must be aware of them. I would recommend this article to designers learning about third party security solutions and technologies. It would also be of interest to someone involved in international medical software applications.
Review #3
Summary of "Securing Embedded Systems for Networks"
http://www.eetuk.com/showArticle.jhtml?articleID=171201927 by David Vaglia
1. TARGET AUDIENCE
This article is intended for the designers of commercial software products that
will be used in embedded systems. This article is written so that designers of
varying levels of expertise can understand it.
But familiarity with the design of embedded systems is assumed.
2. TYPE OF DOCUMENT
This document is organized as a list of helpful hints for the embedded system
designer.
3. SUMMARY OF ARTICLE
The article begins by saying that most systems have priority in making the
product ship on time first, hope the reliability is pretty good and add on
security as an afterthought. The
authors stress that components must be not only be mostly modular but truly
separate. The architecture and
three main considerations of “time-to-market”, “reliability” and
“security” must be designed in from the beginning.
There are a number of tools and standards available to help the designer
achieve his goals. There is a considerable part of the article that explains how
to divide the application into kernel and isolated memory. The authors hope that
the advice given in this article from this will help in the development of
embedded systems.
4. CONCLUSION
The most important idea learned from this article is that all design goals must
be considered from the beginning. Creating a plan that accounts for on time
shipping, reliability, and security will create the most robust product.
I would recommend this for someone who is designing software to be
used in an embedded system environment.
Summary of "Distributed Shared Memory"
http://en.wikipedia.org/wiki/Distributed_shared_memory by David Vaglia
1.TARGET AUDIENCE
This article is intended for individuals seeking a definition of the subject
matter. This article is of a medium technical level and some background in the
areas of networking and distributed systems is recommended.
2. TYPE OF DOCUMENT
This document is organized as an encyclopedia article. It includes a definition
of the main subject, Distributed Shared Memory, and links to related terms.
3. SUMMARY OF ARTICLE
Distributed Shared Memory can be implemented in two ways. The first is as an extension of the operating system and is transparent to the programmer. The second way is to implement it as a library and the user has to be concerned with many underlying issues. This second approach is more portable according to the article. There are several links to samples of DSM, but only one, TreadMarks, was working.
4. CONCLUSION
The most important idea learned from this article is that the choice of
Distributed Shared Memory implementation can have a large impact on application
developers. I would recommend this
article to anyone is planning on working on a DSM or developing software
involving a DSM.
http://en.wikipedia.org/wiki/XML-RPC by David Vaglia
1.TARGET AUDIENCE
This article is intended for individuals seeking a definition of XML-RPC. This
article is of a medium technical level and some background in the areas of
networking and programming is recommended.
Familiarity with data types is helpful in this case.
2. TYPE OF DOCUMENT
This document is organized as an encyclopedia article. This is another article
from Wikipedia. It includes a definition of the main subject, XML RPC, and links
to related terms.
3. SUMMARY OF ARTICLE
The article is a complete definition, specification, sample code and history of the XML-RPC protocol. This article shows how a RPC request can be packaged in XML and sent to a server. The response is read in and the result or error is extracted. The entire protocol is easy to understand and use.
4. CONCLUSION
The most important idea learned from this article is that the Distributed
Technologies can be simple to create and use. The technology is built using the
HTTP protocol and XML. Many useful
functions could be developed quickly using just a few simple tools. I would
recommend this article to anyone is planning on working on an application that
involves any distributed processing between applications.
Summary of The Effect of Latency on User Performance in Warcraft III
by David Vaglia
1.TARGET AUDIENCE
This article is intended for designers of distributed applications. This article
is especially focused on Real Time Strategy games.
2. TYPE OF DOCUMENT
This document is organized as a case study of a single game and the effects of
network latency on that game. There
are introduction, background, test results, analysis, conclusions, etc.
3. SUMMARY OF ARTICLE
The article asks a question about how network latency affects the play for a specific type of game, Real Time Strategy. The article analyzes the different game phases, tracks network traffic, imposes additional stresses by adding network latency and reviews the result from the additional stress. The authors discovered that the games generally produced a steady stream of small data packets that were suitable for enjoyable play even if the network connection was a low speed modem. The key algorithm in this case is that the server synchronized game play on all computers by adding latency to the “fast” computers so that all actions were executed in real time on all computers at the same time.
4. CONCLUSION
The most important idea learned from this article is that there are algorithms
and strategies that can be applied to distributed applications that can minimize
the effects of network latency or “lag”. This can improve the user’s
satisfaction with the application. I would recommend this article to anyone
interested in learning more about internet games and the network protocols
involved and to anyone interested in distributed applications.
Summary of "Algorithms for High Performance, Wide-Area, Distributed File Downloads"
http://www.cs.utk.edu/~plank/plank/papers/CS-02-485.pdf
by David Vaglia
1.TARGET AUDIENCE
This article is intended for designers of algorithms used for downloading files
that are stored on a Wide Area Network. This article is very technical, yet it
is easy to understand for someone who is familiar with network terminology.
2. TYPE OF DOCUMENT
This document is organized as a white paper study of a several algorithms used
for downloading files. There are
introduction, description of Network Storage stack, test results, analysis,
conclusions, etc.
3. SUMMARY OF ARTICLE
The article asks a question about how a file stored in a distributed Wide Area network can best be downloaded and given to an end user. The article gives an introduction to the Network Storage Stack and Logistical Runtime system. This system is a layer model that divides a network storage system into a series of abstract levels. The authors note that the fastest download scheme is to use hundreds of simultaneous TCP connections, but this method can lead to considerable network congestion by exploiting a weakness in network congestion control. The authors limit themselves to one or two connections in an attempt to follow rules of congestion control. The article then describes several download algorithms and experiments used to evaluate the algorithms.
4. CONCLUSION
The most important idea learned from this article is that there are multiple
algorithms and strategies that can be used for downloading files in a Wide Area
Network. There are also ways to fine-tune the algorithms for different
applications such as downloading small or large files or streaming video. I
would recommend this article to anyone interested in learning more about
Internet communication and the algorithms involved and to anyone interested in
distributed applications. This
article also shows the value of having redundant resources in a distributed
application.
Summary of "Windows XP joins embedded system battle"
http://www.itworld.com/Comp/2245/IDG011128XPjoinsbattle/
by David Vaglia
1. TARGET AUDIENCE
This article is intended for designers and users of non-PC computing devices. It is especially focused on the choice of operating systems in those devices.
2. TYPE OF DOCUMENT
This article is an introduction to a new Microsoft product. This is written as a short magazine article.
3. SUMMARY OF ARTICLE
This article begins by noting that there are an ever increasing number of small non
traditional applications for operating systems. Many common devices are having programs embedded in them. Microsoft does not have a large share of this market at this time. The Microsoft operating system is not stable enough to run critical real time devices. However, Microsoft is releasing a new real time operating system. Microsoft is also competitive in the
non-critical device market, especially because of the large number of applications available for the embedded operating system. The article further compares several other operating systems available.
4. CONCLUSION
The most important thing in this article is that choice of operating system is very important for real time critical devices. There are many choices available for non critical devices. I would recommend this article to users who are looking for an introduction to issues concerning embedded systems.
Summary of "Managing Distributed Software Development"
by David Vaglia
http://www.stickyminds.com/sitewide.asp?ObjectId=6002&Function=DETAILBROWSE&ObjectType=ART
1. TARGET AUDIENCE
This article is intended for software developers and managers of software
development teams.
2. TYPE OF DOCUMENT
This article is an introduction to a new type of workplace environment. This is written as a short magazine article.
3. SUMMARY OF ARTICLE
This article begins by noting that software development traditionally was done
in small or large groups that were physically close. Communication was
easy since the others workers were only a few feet away. However, today,
there has been a shift in the environment as workers on the same team are
distributed geographically. The article discusses this problem and
provides solutions and describes the many technologies that can be used to bring
workers together. The article list examples such as web conferencing and
sharing repositories of information and source code.
4. CONCLUSION
The most important thing in this article is that communication is very important for
success in the software development field. There are many tools available to
keep workers up to data with each other. I would recommend this article to
anyone working on a software team that is distributed.
Summary of "Step-by-Step Guide to Distributed File System"
http://www.microsoft.com/windows2000/techinfo/planning/fileandprint/dfssteps.asp
by David Vaglia
1. TARGET AUDIENCE
This article is intended for system administrators with an Active Directory
installation.
2. TYPE OF DOCUMENT
This article is an guide to configuring Distributed File Systems (DFS).
3. SUMMARY OF ARTICLE
This article is an introduction to some of the concepts in Active Directory
service. This article provides detailed configuration steps for an administrator
to use. This article takes the time to explain the settings and the impact
of choosing different configurations. This is very technically detailed
and there is an assumption the user is very familiar with the concepts of Active
Directory and DFS. The article explains how to create shares, mount volumes and
test them.
4. CONCLUSION
The most important thing in this article is that multiple servers can be used to
provide replication and fault tolerance. Using multiple computers will provide
the highest level of services useful for distributed systems. I would recommend this article to
system administrators or anyone wishing to learn more about DFS.