Day 4 At Microsoft SOA and Business Process Conference

Last day at the conference I will following these sessions:

Developing Applications for the Web 2.0 Jon Flanders
Applying Maximum Sustainable Throughput to a Management/Operations Strategy Scott Colestock
Integrate Your Legacy Applications with the BizTalk Adapters for Host Systems and Host Integration Server Paul Larsen, Microsoft
Web Services Software Factory Don Smith Microsoft


Developing Applications for the Web 2.0 Jon Flanders

Here are the notes of this session:

Web 2.0 means nothing

Web 2.0 is really all about services and service usage models

AJAX is not new already there for years (outlook web access), Google use it Google maps and there u go
Web application more as a desktop application

AJAX moving away from XML, efficiency

ATLAS Microsoft’s entry point in AJAX --> Library to build these kinds of applications

Mashup, what does it mean:

A mashup is a website or web application that seamlessly combines content from more than one source into an integrated experience.
(http://en.wikipedia.org/wiki/Mashup_(web_application_hybrid))

Web page or site that gets functionality from services from other sites
Web services bridges help to enable mashups by getting around browser security

Next session this morning is:

Applying Maximum Sustainable Throughput to a Management/Operations Strategy Scott Colestock Trace Ventures, LLC

Dense presentation

Why to apply
maximum Sustainable Throughput (MST)
Performance Goals
Measurement these goals against operational parameters
determine how to scale out

Project phases (5 phases)

Requirement phase
* Establish performance goals
* Identify performance risk area
* Perform first cut estimate size
* Plan sufficient testing

Design Phase
* Acquire detail throughput latency (production usage logs)
* Investigate performance risk mitigations
* Reference system size estimate

Implementation Phase
* Performance test cases
* Build test bed begin
* Validate test

Verification Phase
* Test verify sustainable throughput latency
* Identify bottlenecks adjust hardware

Release Phase
* Verify platform level optimizations
* Setup/back. log shipping, data retention, purging in production

What is missing

going to all phases, transfer operations groups, need to describe what outr systems what operation parameters, need to measure op group

if no operational parameters
look cpu, mem. disk idle %, sql lock wait
BizTalk op param
Pro activate monitoring,

MST, expression performance goals

highest load message traffic in system can handle in a production environ
message/sec transaction/sec
system = application + hardware config
performance testing. database maintenance, SQL agents, bam config, operational queries

Highest system can handle
performance goals drive performance test (latency, batch,
if perf test show system can handle, while meeting goals then basis for performance goals

Three parts expressing performance goals

1) transaction rate msgs/sec during the day (transaction rates or vol as function of time)
2) max acceptable latency, 95% satisfied in 15 sec or less or all requires must return in 2 sec
3) distribution of message types, sizes

Performance goals recap

entiuy often without goals without expressed in clear goals
if you do it you have your operational parameters

Loadgen tsting trans ratets
lactency test completion

Measure transaction rates against time
- BAM or tracking (HAT) MOM
- Business activity for transactions rates (BAM Alert or MOM)
- BizTalk Messages doc recv/sec counterhas instance (decicated hosT)
- Granualarity custom perf counters

Measure monitor transaction rates
- Echo transaction wfetch

Measure Latency
- request-response latency (sec) couter for each host
- BizTalk Tracking Database
- BAM (create duration with view)
- Custom counters in pipeline
- batch volume completion elasped in time constraint (tracking database, doc recv/sec)

DEMO measuring latency
- MOM, perfmon, Loadgen, HAT (quering)


Measure distributed of messages sizes and types
- message types
- sizes, adapters often perf counters
- types and size, tracking database and MOM

Determine scale out strategy

so you have been measuring (limits reached app + hardware)
how to scale out
need to understand BizTalk functions as shock absorbers

BizTalk Host concept
- artifacts container, runs as instance as NT service
- instances on machines

Job host pull message box (tech detail what happens in hosT)
- each has a table hostname Q
- message arrival several actions going to happen
- several actions happen

How does BizTalk acts as an shock absorber

Choosing workloads for new servers
- General counters perf counter spool size counter
- if climbing dig individual host queue count
- reach host queueu refl host not one server
- add server relieve orchestration bottlenecks when host queue grows for processing
- if orchestration times appears been growing cp bound

consider adding server ro relieve send host bottlenecks, with host queue length grows send host
relieve server if recv host bottlenecks
add server when message box general counters, track db size grows
makes sure you have archive/purge strategy in place DTA is the key

Summary

take what u learn performance testing
yield operation parameters
express performance goals clearly
while test 1 and 2, produce measure all
use queue behavior

blog www.traceoftoughts.net
scott@traceventures.net

Integrate Your Legacy Applications with the BizTalk Adapters for Host Systems and Host Integration Server Paul Larsen

Connected systems ;BizTalk, SQL Server --> connectors to other worlds like IBM OS/400, CICS, ...

Composite applications
- integrate multiple systems/apps
- preserve integrity, performance systems, high performance routing IP, SNA API, Optimized iSeries Servers
- solutions based on composite apps, orch composite sol

Session Integrator
- reuse 3270 based programs
- extend 4600 financial apps (cics)
- software development

Host File Web Service

Last Session of the conference:

Web Services Software Factory Don Smith Microsoft

All about communication

Architect -- Developer

Software Factories
Guidance Packages
Helping Communication
Challenges addressed
Factory demonstrated
The 2 Service Factories
Related Resources

Software Fact

- Help build specific kind of app
- incorporate variety guidance content types and form factors
* arch, design guidance, patterns (readable/printing)
* reference implementation (execute)
* application blocks (reusable)
* guidance packages (actionable)
- Enable process

Guiance Package

- Open
* Doc description manual/automate task
* result of automation is exposed
- Configurable
* key configuration setting can be
* automate guidance is defined using xml
- Extensible
* guidance can be tailored to meet specific requirements
- Verifiable
* generated code can be verified for compliance within standard

Helping Communication (snap shot written guidance)

Architect --> Developer --> Visual Studio

Service Factory Contents --> asmx services

- designing message ..
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnpag2/html/servicefactory.asp

Generator for services

ASMX release JUL 2006
WCF release DEC 2006

END OF CONFERENCE