Upload
marcelo-marchi
View
216
Download
0
Embed Size (px)
Citation preview
8/4/2019 Q! for Performance & Capacity Management
1/18
Knowing the core. Smarter Decisions.
Q! for FinanceIT Capacity & Performance Management
8/4/2019 Q! for Performance & Capacity Management
2/18
Knowing the core. Smarter Decisions.
:A e m p re sa Q u a n tita
Nosso objetivo: dar aos nossos clientes toda a plataforma de
Solues Analticas em Tempo Real
para criarem empresas de latncia zero e otimizarem a experincia e retornode seus clientes e acionistas.
Misso: Executar anlises automticas que gerem benefcios econmicos!
reas clientes Marketing, Vendas, Finanas, Mercado de Capitais, Operaes, Controle,
Performance e Riscos Corporativos
8/4/2019 Q! for Performance & Capacity Management
3/18
Knowing the core. Smarter Decisions.
SistemasdeR
eportingeBusines
sIntelligence
Anlise
sautom
ticasc
ombaix
alatnc
iaDecises
emtemp
oreal
Complexidade da anlise
Importnciada automao
Solues Analticas
Performance
Corporativa
Vendas e
Marketing(multicanal)
Operaes e
controle
Mercado
Financeiro ede Capitais
8/4/2019 Q! for Performance & Capacity Management
4/18
Knowing the core. Smarter Decisions.
Q! for Marketing Dynamic Cross and Up-Selling Live Sales Forecast Scores and Propension Behavioral Targeting & Analytical CRM Web Analytics Competitive Intelligence
Q! for Finance Market Screening Financial BehavioralTargeting Algorithmic Trading Fraud detection Risk Management Collection Scoring
Q! for Business Corporate Performance Management Real Time Business Intelligence Business Predictive Modeling
Quantita Servios e Tecnologia
8/4/2019 Q! for Performance & Capacity Management
5/18
Knowing the core. Smarter Decisions.
! Q for Finance Mercado de Capitais
Desempenho requisito para a competio Estratgia do Negcio
Liderana no segmento Rapidez de aes e go-to-market Arbitragens e High Frequency Trading Maiores retornos Reteno e lealdade do cliente
Objetivos Construo de um ambiente e procedimentos analticos
utilizando a tecnologia Modelos de monitorao e determinao da capacidade
corrente dos recursos de TI Modelos de determinao da utilizao mxima dos
recursos de TI e drivers de performance;
8/4/2019 Q! for Performance & Capacity Management
6/18
Knowing the core. Smarter Decisions.
&Performance Capacity Management
8/4/2019 Q! for Performance & Capacity Management
7/18
Knowing the core. Smarter Decisions.
&Performance Capacity Analytics
Alinhamento dos requisitos denegcio
Gesto de nveis de
serviosde TI,previses (forecast) ealinhamento comrequisitos de negcio
Capacidade deServio de Negcio Capacidade deServio de TI Capacidade deComponente de TI
Performance atual dosServios de TI
- Gesto tcnica de equipe
especializada comconhecimento de todos osrecursos de TI
Desempenho de componentesda infraestrutura de TI
- Aes tomadas por equipes de
suporte e manuteno doambiente
Performance & Capacity Analytics-Anlise da influncia: relao da demanda por recursos computacionais e o uso
destes recursos-Application sizing: estimativas de recursos para suportar os nveis de serviosde cada aplicao-Modelagem: Previses do comportamento dos Serviosde TI dado um volumede trabalho. Principais tcnicas: anlise de tendncia, modelos analticos,simulaes
8/4/2019 Q! for Performance & Capacity Management
8/18
Knowing the core. Smarter Decisions.
Processo de Gesto (contnua) da Capacidade da BVMF foco na capacitao e acompanhamento da performance de TI Componentes de hardware: PCs, servidores de arquivos, de web, de banco de dados,
mainframes
Equipamentos de rede: LANs, WANs, bridges, routers
Perifricos: dispositivos de armazenamento, impressoras
Softwares: sistemas operacionais e sistemas de rede, pacotes de software internos ouadquiridos
BDC Banco de Dados de Capacidade
Capacidade deServio de
Negcio
Capacidade deServio de TI
Capacidade deComponente de
TI
BDC Banco deDados de Capacidade
Planode
Capacidade
Gesto da Demanda
8/4/2019 Q! for Performance & Capacity Management
9/18
Knowing the core. Smarter Decisions.
Potential to cause all the work to be scrapped
Huge negative consequences of performancefailure Damaged customer relations Loss of business Loss of credibility Cost of delay & additional resources Performance defect identified during
maintenance stage could cost 50 to 100 timesmore to fix as it would have cost during earlystage of development lifecycle.
! :Q for Finance os custos da baixa performance
8/4/2019 Q! for Performance & Capacity Management
10/18
Knowing the core. Smarter Decisions.
Process by which software is tested and tuned
with the intent of realizing the requiredperformance
Performance categories
Speed Scalability Stability
Performance engineering helps IN Increase revenue by ensuring optimum system performance
Optimize provisioning & utilization of infrastructure Improve availability by resolving performance & scalability Reduce maintenance costs Avoid system failure requiring scrapping & writing off development
effort Self impressive facts about the release
! :Q for Finance Engenharia de Performance
8/4/2019 Q! for Performance & Capacity Management
11/18
Knowing the core. Smarter Decisions.
Performance testing & sizing
Performance Regression Test, System Load testing, SystemBenchmarking
Performance tuning and optimization Optimal tuning guidelines for production setup JVM, Pool, logs,
App/Web server, DB, OS, etc.
Effectively applying skills, technologies and tools System diagnostics End to end profiling, recommendations for overall system
scalability
Capacity planning Determine expected production capacity, Facilitate capacity
management Performance Centric development Client side performance engineering
! :Q for Finance Servios de Engenharia dePerformance
8/4/2019 Q! for Performance & Capacity Management
12/18
Knowing the core. Smarter Decisions.
Performance Engineering Strategy
8/4/2019 Q! for Performance & Capacity Management
13/18
Knowing the core. Smarter Decisions.
Challenges in Performance Testing
Performance engineering awareness
Right tool selection Scripting expertise Physical/practical resource limitations on
number of human users
Appropriate Hardware setup Governing the behavior of simulated users
according to a set of policies Consistent, repeatable actions
Accurate measurement of response times Generation and comprehensive analysis of
results
8/4/2019 Q! for Performance & Capacity Management
14/18
Knowing the core. Smarter Decisions.
Features in an Ideal Load Test Tool
Support for popular protocols viz. Http, Https,
AJAX, WAP, SMTP etc Integrated resource monitoring Simple and intuitive User interfaces and wizards Easy to use parameterization features
Real time progress of the test runs Support for high concurrency of virtual users Sophisticated and comprehensive reporting and
analysis capabilities
8/4/2019 Q! for Performance & Capacity Management
15/18
Knowing the core. Smarter Decisions.
&Tools Utilities
Performance Testing , , , , , ,LoadRunner SandStorm Silk Performer Jmeter Grinder Webload,OpenSTA
Web Servers Apache IIS
Application Servers Apache Tomcat Jboss
WeblogicWebsphere
Databases Oracle MS SQL Server My SQL PostgreSQL
System Profiling , , , , ,HP Deep diagnostics Jprobe DynaTrace Jprofiler YourKit
InfraRed , ,GCViewer HttpSniffer
8/4/2019 Q! for Performance & Capacity Management
16/18
Knowing the core. Smarter Decisions.
Case I
Background Recognized market and thought leader in Workforce
management -Delivers Innovative solutions and solve complex workforce
.related issues
, + .30 million users per day in 30 000 hiring locations
+ :500 Fortune 1000 clients includes the largest consumer, , , ,goods automotive telecom customer service payroll
.services
Challenges %90 of customers use Applet version of the application
Support for load testing of the applets and java objects
Affordable proof of concepts for pre sales activities
8/4/2019 Q! for Performance & Capacity Management
17/18
Knowing the core. Smarter Decisions.
Accomplishments
Defined Performance testing strategy and process
Helped Client by setting up Test Lab with isolated(LAN Completely free from other corporate).network activity Setup included : . , / : , ,9 servers Dual 3 2 Ghz PentiuApp Web servers IIS Jboss, Weblogic Apache
: , & , & ,Database servers Oracle9 10 11 SQL server2005 2008
, ,m 4 2 GB RAM 500 GB HDD
+15 Agent machines for launching Virtual users
SandStorm support for load testing of applets andjava objects
Tested application with 5k concurrency levels Dedicated support for deployment and customization Improved customer satisfaction because of onsite
benchmarking
8/4/2019 Q! for Performance & Capacity Management
18/18
Knowing the core. Smarter Decisions.
Case 2
Background Leading telecommunication solutions and services provider Instant messaging solutions for mobile devices Innovative solutions for mobile and internet industry
Challenges
Load test product based on OMA IMPS mobile protocol Simulate concurrency of 25k online users