Capacidade Planejada Rodrigo Albani de Campos - [email protected] - @xinu Wednesday, August 25, 2010
Capacidade Planejada
Nov 11, 2014
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Agenda
• Motivações
• Capacidade + Velocidade
• Universal Scalability Model
• Extra: Dimensionamento de pools
Wednesday, August 25, 2010
Impacto na imagem do produto
Wednesday, August 25, 2010
Impacto na imagem do produto
Wednesday, August 25, 2010
Amazon 2010Q1 Net Sales$7.13 Billion Dollars
SLA Horas / Trimestre Impacto $
98% 43,2 $142.600.000
99% 21,6 $71.300.000
99,90% 2,16 $7.130.000
99,99% 0,216 $713.000
99,9990% 0,0216 $71.300
99,9999% 0,00216 $7.130
Wednesday, August 25, 2010
ReceitaImpacto de lentidão na entrega de conteúdoThe User and Business Impact of Server Delays, Additional Bytes, and HTTP Chunking in Web Search - Eric Schurman (Amazon), Jake Brutlag (Google)
Distinct Queries/User
Query Refinement
Revenue/User
Any Clicks Satisfaction Time to Click (increase in
ms)
50ms
200ms
500ms
1000ms
2000ms
0 0 0 0 0 0
0 0 0 -0,30% -0,40% 500
0 -0,60% -1,20% -1,00% -0,90% 1200
-0,70% -0,90% -2,80% -1,90% -1,60% 1900
-1,80% -2,10% -4,30% -4,40% -3,80% 3100
$85.000.000 em três meses na Amazon
Wednesday, August 25, 2010
ReceitaImpacto de lentidão na entrega de conteúdoThe User and Business Impact of Server Delays, Additional Bytes, and HTTP Chunking in Web Search - Eric Schurman (Amazon), Jake Brutlag (Google)
Wednesday, August 25, 2010
“Fast isn’t a feature, fast is a Requirement”
Jesse Robins - OPSCode
Wednesday, August 25, 2010
Composição de páginas na Web - Top SitesWeb Metrics: Size and number of resources - Sreeram Ramachandranhttp://code.google.com/speed/articles/web-metrics.html
Average # of resources /
page
Average # of hosts / page
Average document size
Kb
42,14 8,39 477,26
Wednesday, August 25, 2010
http://www.webpagetest.org/ TEST RESULTSJuly 2010
Load Time
Mercado Livre 1,972Ebay 1,999Amazon 4,777newegg 6,848Bestbuy 7,508Submarino 10,436Casas Bahia 15,09
First Byte Start Render
Complete Time (s)
Requests Bytes In
329 ms 543 ms 2,491 22 74 KB
399 ms 1493 ms 3,103 36 237 KB
504 ms 1105 ms 6,289 69 454 KB
328 ms 1211 ms 7,891 138 459 KB
447 ms 1733 ms 10,41 99 676 KB
250 ms 2474 ms 10,436 151 1,125 KB
500 ms 4401 ms 15,799 100 732 KB
Wednesday, August 25, 2010
Capacity and Velocity
Wednesday, August 25, 2010
Queuing Theory 101
Service time: Tempo de ocupação do recurso (s,ms,μs)
Arrival rate: Taxa de chegada de requisições para o recurso (hit/s,qps,etc...)
Little’s Law: The long term average number of customers in a stable system L is equal to the long term average arrival rate λ, multiplied by the long term average time a customer spends in the system, W
Wednesday, August 25, 2010
Average Service Time against Arrival Rates
0
200
400
600
800
1000
1200
1400
1600
0
0,01
0,02
0,03
0,04
0,05
0,06
0,07
0,08
0,09
0,1
0 50 100 150 200 250 300
Freq
uency
Service Time (s)
Hits/s
Service A
Service Time Frequency
Wednesday, August 25, 2010
Average Service Time against Arrival Rates
0
1000
2000
3000
4000
5000
6000
0
0,5
1
1,5
2
2,5
0 20 40 60 80 100 120 140 160 180 200
Freq
uency
Service Time (s)
Hits/s
Service B
Service Time Frequency
Wednesday, August 25, 2010
APDEX - http://www.apdex.org
0
1000
2000
3000
4000
5000
6000
0
0,5
1
1,5
2
2,5
0 20 40 60 80 100 120 140 160 180 200
Freq
uency
Service Time (s)
Hits/s
Service B
Service Time FrequencySatisfied
Tolerating
Frustrated
Wednesday, August 25, 2010
APDEX - http://www.apdex.org
Satisfied Tolerating Frustrated
Wednesday, August 25, 2010
APDEX - http://www.apdex.org
SamplesSatisfiedTolerating
Apdex
1200010000
800
86,67%
Wednesday, August 25, 2010
Forecasting (Sort of...)
Wednesday, August 25, 2010
Forecasting (Sort of...)
0
5
10
15
20
25
30
35
40
0 10 20 30 40 50 60 70
Throughput
Virtual Users
System C
Measured X
Measured ThroughputUsers (N) X(N)
1 2,912 5,674 10,868 18,65
16 25,9132 36,6864 37,34
Wednesday, August 25, 2010
Forecasting (Sort of...)
0
5
10
15
20
25
30
35
40
45
0 10 20 30 40 50 60 70 80 90
Throughput
Virtual Users
System C
Measured X Poly. (Measured X)
Wednesday, August 25, 2010
Forecasting (Sort of...)
0
5
10
15
20
25
30
35
40
45
0 10 20 30 40 50 60 70 80 90
Throughput
Virtual Users
System C
Measured X Poly. (Measured X)
Wednesday, August 25, 2010
0
5
10
15
20
25
30
35
40
0 20 40 60 80 100 120 140
Throughput
Virtual Users
System C
Measured X Modeled X(N)
Using the Universal Scalability ModelNeil J. Gunther http://www.perfdynamics.com/Test/gcaprules.html#sec:scalability
Wednesday, August 25, 2010
0
10
20
30
40
50
60
0 100 200 300 400 500 600
Throughput
Virtual Users
System D
Measured X Modeled X(N)
Using the Universal Scalability ModelNeil J. Gunther http://www.perfdynamics.com/Test/gcaprules.html#sec:scalability
Wednesday, August 25, 2010
Conclusões e Considerações
• Velocidade é tão importante quanto disponibilidade
• Fast is a requirement
• O SLA deve ser definido considerando a experiência do usuário
• Não existem bolas de cristal
Wednesday, August 25, 2010
In God we trust.Everyone else
please show me the data.
Wednesday, August 25, 2010
http://capacitricks.wordpress.com/
Wednesday, August 25, 2010
How many servers do we need ?
Wednesday, August 25, 2010
Redundância
55%
55%
55%
55%
Wednesday, August 25, 2010
Redundância
73,3%
73,3%
73,3%
X
Wednesday, August 25, 2010
Redundância
110%
110%
XX
Wednesday, August 25, 2010
How many servers do we need ?
Wednesday, August 25, 2010
How many servers do we need ?
Wednesday, August 25, 2010
Perguntas ?
Wednesday, August 25, 2010
Related Documents
