december 9, 2002 unicon / ibs: adam rybicki michael erdely sun microsystems: jeff weiss testing and...

17
December 9, 2002 UNICON / IBS: Adam Rybicki Michael Erdely Sun Microsystems: Jeff Weiss Testing and Proving uPortal’s Scalability at the Sun iForce Center

Upload: magnus-wilson

Post on 24-Dec-2015

217 views

Category:

Documents


0 download

TRANSCRIPT

December 9, 2002

UNICON / IBS:Adam RybickiMichael Erdely

Sun Microsystems:

Jeff Weiss

Testing and Proving uPortal’s Scalability at the Sun iForce Center

• Test Objectives• About iForce Centers• Content Selection Criteria• Test Environment• How We Tested?• Results• Conclusions• Questions

Agenda

• To prove that the uPortal framework does not prevent scalability

• NOT to prove that every channel will scale the same way

• uPortal represents J2EE’s presentation layer—not the business logic layer

Test Objectives

Sun’s iForce Center Mission

To addresses customer business objectives by helping to reduce risk and speed time to implementation of solutions that help lower their total cost of ownership and increase the productivity of their assets

Sun’s iForce Centers

• Assessments• Evaluate your business processes • Identify areas for improvement.

• Solution Demonstrations• Hardware and software products for a

custom demos• Proof-of-Concepts (PoC)

• Prototype business solutions matched to your needs.

• Test to see if proposed solution is viable.

• Sizing and Tuning• Together we will spec out the optimal

configuration of hardware and software to serve your business.

• http://www.sun.com/executives/iforce/

• Select applications “heavy” in presentation and “light” in business logic

• Access to external data eliminated• Mix applications that have business

logic with those that only present data

• Applications with business logic should utilize back-end servers

• No way to simulate legacy systems• No application accessed the database

server, but frequent logins simulated that type of application

Content Selection Criteria

Test Environment

Sun ONE Product Stack

uPortal Server

Community

Data

Applications

Transactions

Sun ONEApplication and

Web Servers

Sun ONEIntegration

Server

Services

Sun ONE Directory Server

Reports

InformationAssets

Su

n O

NE

Stu

dio

SolarisTM Operating Environment

Sun ONECommunication Services

Mail CalendarStarOffice

CustomBusiness Logic

• Database was preloaded with 25,000 users• A simulated session consisted of:

• Visit the initial page• Logon• Visit every tab• Interact with Webmail and Calendar• Logout

• Every container was “primed” by manually logging on and visiting every tab

• 10 new virtual users added every minute per server

• Random delays (0-10 seconds) between simulated clicks

• Tests ended after the response time was over 10 seconds

How We Tested

0

5

10

15

20

25

10

70

13

0

19

0

25

0

31

0

37

0

43

0

49

0

55

0

61

0

67

0

73

0

79

0

85

0

91

0

97

0

10

30

0

20

40

60

80

100

120

140

Page Time Hits Per Second

Load size 10 per. Mov. Avg. (Page Time)

15 per. Mov. Avg. (Hits Per Second)

0

5

10

15

20

25

10 70 130

190

250

310

370

430

490

550

610

670

730

790

850

910

0

20

40

60

80

100

120

140

Page Time Hits Per Second

Load Size 10 per. Mov. Avg. (Page Time)

15 per. Mov. Avg. (Hits Per Second)

1 Tomcat 1 Sun ONE

Results

05

10152025303540455055606570

10

15

0

29

0

43

0

57

0

71

0

85

0

99

0

11

30

12

70

14

10

15

50

16

90

18

30

19

70

21

10

22

50

23

90

0

100

200

300

400

500

600

700

Page Time Hits Per Second

Load Size 10 per. Mov. Avg. (Page Time)

15 per. Mov. Avg. (Hits Per Second)

05

101520253035404550556065

10

13

0

25

0

37

0

49

0

61

0

73

0

85

0

97

0

10

90

12

10

13

30

14

50

15

70

16

90

18

10

19

30

20

50

21

70

0

100

200

300

400

500

600

Page Time Hits Per Sec

Load Size 10 per. Mov. Avg. (Page Time)

15 per. Mov. Avg. (Hits Per Sec)

Results

4 Tomcats 4 Sun ONEs

05

101520253035404550556065707580859095

100

10

16

0

31

0

46

0

61

0

76

0

91

0

10

60

12

10

13

60

15

10

16

60

18

10

19

60

21

10

22

60

24

10

25

60

0

200

400

600

800

1000

1200

1400

1600

Page Time Hits Per Second Load Size

10 per. Mov. Avg. (Page Time) 15 per. Mov. Avg. (Hits Per Second)

05

101520253035404550556065707580859095

100105110115120

10

15

0

29

0

43

0

57

0

71

0

85

0

99

0

11

30

12

70

14

10

15

50

16

90

18

30

19

70

21

10

22

50

23

90

0

200

400

600

800

1000

1200

1400

1600

Page Time Hits Per Second Load Size

10 per. Mov. Avg. (Page Time) 15 per. Mov. Avg. (Hits Per Second)

Results

8 Tomcats 8 Sun ONEs

• THESE RESULTS ARE VALID ONLY FOR THIS SPECIFIC CONFIGURATION

• In contrast to previous versions, 2.0.2 is very stable and much more scalable

• Database must be well-tuned• More throughput could be obtained

from Sun ONE servers, if we spent time tuning them

• uPortal and Sun ONE servers put more load on the Sun ONE LDAP server than expected

Conclusions

• uPortal pages with many channels (more than 10 on a page) create too many threads

• We should work on improving the logon processing

• Web containers were not intended to house applications, but their presentation only

Conclusions

Conclusions

Conclusions

[email protected]

[email protected]

[email protected]

http://www.uportal.biz

http://www.sun.com/executives/iforce/centers/

http://www.sun.com/software/index.html

Questions?