|
Past Meetings
May 8, 2002
Demystifying Service
Oriented Architectures (SOA) and Web Services - Jeff Buzen
We are pleased to announce that Jeff Buzen, past CMG President and
frequent speaker at Boston
CMG, will present hot topics in the area of Web Services. Jeff's first
presentation will examine Web services architecture. This presentation
will explain the basic standards that form the foundations of Service
Oriented Architectures and Web Services, including SOAP, XML, WSDL, and
UDDI.
------------------------------------------------------------------------------------------------------------
Measuring and Managing the Performance of Web Services - Jeff Buzen
Jeff's second talk will present key performance measurement and
management issues associated with Web Services, describe various options
for the development and deployment of measurement tools, and analyze both
the difficulties and the advantages of adapting established performance
management technologies to the new Web Services environment.
------------------------------------------------------------------------------------------------------------
Workload Characterization and Capacity Planning for DB2 Universal
Database - Yefim Somin, BMC.
We will have a presentation from Yefim Somin, BMC Software, on the
performance management of DB2 UDB databases. DB2 UDB is a complex,
multi-tier, distributed database with parallel execution features. To
track resource use and performance of many applications and users putting
a load on this environment requires acquisition and processing of data
from different sources. This paper describes architectural features of
DB2 UDB relevant for workload analysis and an approach to producing
workload characterization for performance management and capacity
planning.
Nov. 14, 2002
Speaker: Ned A. Diehl,
The Information Systems Manager, Inc.
Virtual tape (VT) subsystems are an important part of a large
percentage of OS/390 installations. While, they do a beautiful job of solving
many traditional tape problems, they are not all things for all
installations. VT systems present some interesting challenges, since
while they look like real tape drives to standard OS/390 performance
monitors, there are important internal activities that those monitors do
not see. The two primary implementations (IBM VTS and STK VSM) both
produce SMF records which provide metrics unique to virtual tape.
Approaches to analysis will be discussed and will include both an RMF
view and virtual tape measurements. Graphical examples will be drawn from
a variety of different types of installations. While most examples will
be based on STK VSM, specific IBM VTS metric references will also be
provided.
------------------------------------------------------------------------------------------------------------
10:50am -- The Software Pricing Revolution - Release 2.0
Speaker: Jeff Buzen, CMG Past President
The software industry is struggling to devise pricing mechanisms that
provide adequate revenue for vendors while being fair to customers and
competitive across platforms. In an attempt to meet all these goals, IBM
announced an innovative new pricing strategy (sub-capacity pricing under
VWLC) in October 2001. Before this new mechanism could be widely adopted,
it was radically restructured through a second announcement on April 30,
2002. This talk examines the basic philosophy behind the new pricing
mechanisms and describes some important - and possibly unintended -
implications for optimizing cost/performance on IBM's zSeries platforms.
------------------------------------------------------------------------------------------------------------
1:30pm -- Distributed
Server Consolidation: What are the Performance Trade-offs ?
Speaker: Mike Matchett, Senior Product Marketing Manager, BMC
The drive to lower IT Total Cost of Ownership (TCO) leads to
distributed system server consolidation and shared environments. New
business trends including quickly expanding e-commerce web servers also
steer IT towards larger, centralized platforms. These trends present a
host of challenges that aggressive performance management best practices
can address: planning how to best consolidate workloads, optimizing
transaction performance while minimizing cost per transaction, purchasing
the right resources at the right time, evaluating disaster recovery
fail-over scenarios, and supporting fair cost allocation (chargeback). By
retaining clear insight into "who" is doing "what" in
consolidated hosts and proactively planning for change with predictive
modeling, companies can save money while delivering optimal guaranteed
service.
------------------------------------------------------------------------------------------------------------
3:00pm -- Application
Tuning: What can you do when the problem is in the system ?
Speaker: Sharon Antia, Compuware
The performance of CICS applications often depends on resources that
are part of the underlying system. Ever wonder what causes CICS to spend
so much time in DFHSMCK, or why DB2 spends so much time waiting in
DSNB5RDP? What could be the purpose of something called CEEV#GTS...? And
why would your application get locked in by a resource such as IXLR1LCK?
And, while we are on the subject, what is a MUTEX lock? If you have ever
struggled with system resources and how to configure your application to
spend less time exercising them, come join our discussion.
--------------------------------------------------------------------------------
May 9, 2002
iSCSI or how to do
Storage over IP.
We are pleased to announce that Irwin Kraus, frequent speaker at Boston CMG, will
lead off with a presentation on a hot topic in the area of storage area
networks. iSCSI is the Internet SCSI (Small Computer System Interface), a
new Internet Protocol (IP)-based storage networking standard for linking
data storage facilities used to facilitate data transfers over intranets
and to manage storage over long distances. Irwin Kraus, now with Legato,
is well known as a frequent speaker at regional and national CMG meetings,
and has a lot of information to convey on this important topic.
------------------------------------------------------------------------------------------------------------
Parallel Sysplex Tuning: Recent Developments
Joan Kelley, of IBM, will give us an update the latest developments on
tuning parallel sysplex environments. Joan is a frequent speaker at CMG
and other user groups, and she is a well known authority on mainframe
performance, especially on what all those measurements really
mean, and what you can do about it.
------------------------------------------------------------------------------------------------------------
Linux on the Mainframe - Performance Considerations
We will have a presentation from Joe Temple, IBM, on the performance
implications of using mainframe Linux as a high-end platform for
deploying Unix applications. This area has been receiving wide attention
over the past year and there are many interesting and important issues to
consider, especially relating to performance trade-offs between mainframe
Linux and other Unix environments.
----------------------------------------------------------------------------------------------------------
Distributed Storage Tutorial - What Does it All Mean ?
Jaqui Lynch, of Circle4 Consulting, will give us a guided tour through
the growing maze of today's distributed storage landscape. Approaches
such as SAN and NAS are changing how storage is architected and accessed,
and is introducing many network-like elements into the picture. This was
a popular session at CMG2001.
------------------------------------------------------------------------------------------------------------
Nov. 8, 2001
------------------------------------------------------------------------------------------------------------
Linux and USS on zOS: What, When, Why?
Speaker: Robert Johnson, Landmark Systems Corporation
How is Unix Systems Services different from Linux on the mainframe?
What are the performance and capacity metrics available or planned? What
workloads should you be running or thinking of running on these two
platforms? Where should you run your WebServer applications? As you
decide these questions, you will need to compare and contrast the
architectures of the two UNIX models and prepare yourself to use the right
one.
------------------------------------------------------------------------------------------------------------
A Look at DB2 and OS/390 Workload Manager
Speaker: Edward Woods, Candle Corporation
OS/390 Workload Manager which handles system resource allocation and
prioritization for OS/390 subsystems and applications. DB2 is one of the
major exploiters of Workload Manager on OS/390. This session will review
the major constructs and components of Workload Manager and will then
look at how DB2 exploits the facilites of Workload Manager. This will
include a discussion of DB2 distributed applications, and application
environments.
------------------------------------------------------------------------------------------------------------
Edge-side Options for Enhancing Web Site Scalability
Speaker: Dr. Jeffrey P. Buzen, CMG President
Content delivery networks featuring servers at the edge of the
Internet have achieved impressive levels of performance in a number of
real world situations. Some content delivery providers have made bold
claims about the value of their patented rocket science algorithms. This
session provides an independent look at these algorithms and examines
cases where they are, and are not, likely to deliver the performance
benefits they promise. This presentation also discusses the pros and cons
of Web caching, which can be viewed as a grassroots alternative to top
down content delivery services.
------------------------------------------------------------------------------------------------------------
Taking Care of eBusiness: Quality Control of Large-Scale
Web-Based Applications in Real-Time
Speaker: Kenneth J. Harper, CEO and Chief Technical Architect, eBusiness
1, Inc.
This is a case study of the real-time monitoring and management
solution built for the JCPenney.com Internet store. Included are
methodologies for statistical process and quality control of system
services, application components, end-user response time and business
intelligence metrics such as browser-to-buyer ratios, transaction volumes
and revenue streams. The goal is to enable the attendee to integrate
quality control practices into their own management solution thereby
improving overall quality of service to their end-user community .
------------------------------------------------------------------------------------------------------------
Nov. 9, 2000
Key to the success of
Web-based applications is the "end-to-end" response time
perceived by the user. Our program will begin with Marilyn Kanas of
Keynote Systems who will examine the requirements for "Web Load
Characterization" as a necessary component of Web Site Capacity
Planning.
The heart of any eBusiness application is the server. In this preview
of a late-breaking CMG 2000 paper, Dr. Jeff Buzen, CMG National Chairman,
and Dr. Annie Shum - both from BMC Software - will examine some
performance impacts of the just-announced new "Freeway" eServer
z900 64-bit mainframe systems.
Service Level Agreements are increasingly used to govern the
relationship between users and service providers. Mike Tsykin of Fujitsu,
Australia will present his paper on actual implementation of an automated
service level reporting system for a very diverse multi-platform
environment that includes mainframes, Unix, and NT with many varieties of
databases including Oracle, DB2, and SQL Server.
Together, Web and application servers constitute major components of
eBusiness applications. Dr. Jim Bouhana of Performance International will
examine production reporting for Web-based applications based on Web
server logs and application logs for Unix and NT systems.
April 6, 2000
Our Featured Speaker was Dr. Jeff
Buzen, the new National CMG Chairman. Jeff presented his 1999 CMG paper
"Perspectives on Performance Measurement" and also addressed
some of the upcoming challenges facing the CMG organization today. Jeff
is an internationally known scientist and CMG speaker and is an A.A.
Michelson award winner for outstanding professional contributions to
computer performance evaluation.
The Willian J. Mullen Award is presented each year to a CMG paper that
is distinguished for both technical excellence and outstanding
presentation style. The 1999 Mullen Award went to Craig D. Hanson and Pat
V. Crain from Cabledata for their paper "User and Computer
Performance Optimization - A New Model for Efficiency" This paper
addresses the following: "users urge us to optimize their efficiency
and experience - not simply maximize the computer's performance."
Saqib Syed of BMC Software presented his CMG99 Best Paper:
"Capacity Planning for MS Exchange Applications." This
presentation addresses capacity for large-scale messaging by bringing
together measurements from systems and applications areas, considering MS
Exchange architecture, and facilitating "what-if?" questions.
Jean-Pierre Garbani is Director of Capacity Planning and Forecasting
at John Hancock. His presentation "A Roll-Your-Own Capacity Planning
system for a Major
Call Center
Application" provided hands-on perspective on the tasks required to
do capacity planning and forecasting for a complex, enterprise-class
distributed system without fancy monitoring tools.
November 4, 1999
Some of today's hottest performance
issues concern providing fast response time for key applications, with
access over the Web. This requires hardware selection, systems and
applications tuning, and fast Internet communications. In our exciting
Nov. 4th. meeting, we examined various pieces of the overall
"performance puzzle" including OS/390 Applications Tuning,
Coupling Facility performance, a special tutorial on Web architecture and
performance, and the impact of Storage Area Networks.
Joan Kelley, IBM, presented "Coupling Facility Options -
Performance Considerations." As size and speed of processors
increase and coupling technology advances, there are an ever increasing
number of options for coupling facility processors, links and storage and
the way they are apportioned. This presentation looks at all the coupling
facility options which are currently available, situations to which they
are best suited (or not suited), and the performance implications. RMF
reports of various configurations will be used as examples.
Irwin F. Kraus, Amdahl Corporation, presented: "Fibre Channel and
Storage Area Networks: a Tutorial." Storage Area Networking is an
exciting technology that promises to revolutionize storage configuration
and management for Open Systems by delivering greater connectivity,
performance, scalability, and reliability. Like many new technologies,
however, SAN is currently the subject of a great deal of hyperbole. This
paper describes the evolutionary development of SAN and clearly
delineates current practice from long range potential. Analogies to
System/390 are used wherever possible. Fibre channel loop and fabric
(switch) technology is clearly explained.
Nathan Matt, Programart presented " OS/390 Subsystems &
Components: Impact on Application Performance." OS/390 has many
valuable facilities that can improve application performance. However,
applications can suffer when these facilities are under-used or misused.
During this session the speaker will discuss configuring and optimizing
facilities such as: CICS Transaction Isolation, VSAM System Managed
Buffering, Language Environment , and Virtual Lookaside Facility. STROBE
reports will be used to illustrate resource requirements and optimization
techniques. This session is of interest to anyone concerned with
application performance.
Irwin F. Kraus, Amdahl Corporation, presented: The "Internet,
Java, and Network Computing for Mainframers" The Internet is growing
exponentially and has become a common presence in everyday IT life. This
presentation delves into the major forces involved in commercial data
processing on the web. What is Java? What are JavaScripts, applets,
servlets, Beans vs. Java applications? What is a network computer, how is
it different from a thin client? What are the performance implications of
web-based computing? We present an overview of TCP/IP as that is the
enabling glue for today's Internet applications.
May 6, 1999
In this meeting, we featured two of the
Best Paper Award winners from CMG98.
Donald R. Deese, Computer Management Sciences, Alexandria, VA
presented "Evaluating and Improving CICS Performance in MVS Goal
Mode." This paper, winner of a Best Paper award at CMG98, describes
the information available in Goal Mode and provides suggestions about how
to improve CICS performance based on the information provided. Don is a
frequent CMG speaker and is an A.A. Michelson award winner for
outstanding professional contributions to computer performance
evaluation.
Irwin F. Kraus, Amdahl Corporation, presented: "The RMF Type 70
Record - A Plethora of Possibilities." This paper, winner of a Best
Paper award at CMG98, describes the derivation and meaning of LPAR
(Logical Partition) and MVS busies for various logical processor states.
We will also look into how LPAR utilization and configuration information
can be used to accomplish basic LPAR capacity reporting. Irwin is a
frequent speaker at regional and national CMG meetings and is highly
regarded for his expertise.
Steven Roberts, Senior Systems Manager / Database Administrator, John
Hancock Insurance, presented: "Introduction to Oracle(R) Performance
and Tuning." Today, Oracle is one of the most widely used databases
on OS/390, Unix and NT. The effective use of Oracle as a database or
application server requires careful attention to performance, especially
for online transaction processing. This presentation examines Oracle
structure and operation with a view toward performance management and
tuning for high performance applications.
A Performance Round Table discussion was held covering current issues,
especially Y2K preparation and readiness.
November 5, 1998
Jaqui Lynch of
Boston College
(and next year's CMG 99 General Chairman!) presented "Designing High
Performance Web Pages." As usage of WWW technology continues to grow
explosively, more and more attention must be paid to Web performance.
This paper addresses many important issues including Web server
performance, using cache, and the Top 10 performance mistakes in Web
Design.
CMG A.A. Michelson Award winner Dr. Jeff Buzen of BMC Software spoke on
"Modeling Goal Mode in MVS/ESA." Jeff has been working with a
number of distinguished scientists including Dr. Ethan Bolker, Dr. Annie
Shum and Boris Ginis to understand whether a specific set of response
time goals can be met and, if so, how to find priority assignments that
will achieve them.
Sam Nokes, a software consultant for BMC Software Services, previewed his
CMG98 paper entitled "Estimating Network Traffic for TCI/IP
Distributed Application Workload Modeling." Network traffic volume
and workload characterization are key steps in capacity planning for
distributed applications. This paper describes a methodology for
estimating TCIP/IP traffic generated by a distributed application and use
of a queueing model to predict the impact of future traffic growth.
Bill Zahavi of EMC Corporation presented "The World According to a
Storage System." This paper presents requirements - and solutions -
for performance data in situations where storage systems such as
Symmetrix are accessed by multiple hosts simultaneously. Workload
characterization issues are addressed.
June 25, 1998
Getting and Reporting Distributed
Performance Data, by
Lin Merritt, Trigon Blue Cross Blue Shield
Capacity Planning for IMS Parallel Sysplex Data Sharing by
Bryant Osborn, NationsBank
Business Continuity Through the Year 2000, by
Perry Harris, Yankee Group
CICS Tuning Methodologies, by
Noah Perlman, OptiSystems Solutions
March 26, 1998
Irwin Kraus, Amdahl
Characterization of Late Twentieth Century Workloads: CICS/DB2,
CICS/VSAM, Oracle
Dr. Bernie Domanski, BGS Systems
Capacity Planning for Distributed Environments
Dr. Subhash Agrawal, BGS Systems
Measurement and Analysis of Disk I/O in UNIX Environments.
Sudha Jamthe, BGS Systems
Performance Issues of Designing a High Performance Intranet
October 29, 1997
Bob Rogers, IBM "OS/390-MVS Release 3 and Release 4:
Enhancements and New Features." This presentation provides an update
of recent OS/390-MVS features, with emphasis on performance issues.
Dave Beulke, Bell
Atlantic "Sysplex and DB2 Data Sharing." This paper takes you
through the steps of planning and implementing Data Sharing in DB2
Release 5. It won the Mullen award for best presentation at International
CMG in San Diego.
Dan Schutzman, Emerald Software "Automating Application
Performance Management." This presentation outlines an approach to
automating application tuning to help conserve CPU time in MVS mainframe
environments. A case study will be presented.
Michael Stadelmann, BGS Systems "Measuring End-to-End Response
Time on the Web." This paper presents techniques for measuring
end-to-end response time perceived by a user browsing Internet or
Intranet Web applications. These techniques can be applied to any Web
browser enabled for Javascript, regardless of platform.
June 26, 1997
...Don Deese, President, Computer Management Sciences Evaluating
CICS Version 4.1 Performance under the Workload Manager
...Joan Kelley, IBM Parallel Sysplex Update: Recent Experiences with
DB2, VSAM and other Coupling Facility exploiters.
...Bill Besenyei, Amdahl Thin Clients: Evaluating Three Major Alternatives
...Jaqui Lynch, Boston
College UNIX
Performance Management: It Doesn't Have to Cost a Fortune
March, 1997
...Steve Samson, Candle "Evolution of MVS Workload
Performance Management"
...Ed Saulnier, Coolidge Systems "The Year 2000 Bug: It Could be
a Bigger Problem Than You Think!"
...Perry Harris, Yankee Group "Year 2000 Compliance: A Year-End
Review and Status Test"
...Gene LeGanza, Programart"The Hidden Risk in Year 2000
Implementations"
Go back
to GBCMG home page
|
