finalTBA Benchmark Testing report final April 2004

Administrator

Description

TBA Web Server Installation & Performance Testing 1 TAFE QUEENSLAND Videolinq During February 2004, Integrated Vision undertook a major upgrade of the hardware infrastructure of the ‘To Be Announced’ (TBA) web server that provides conference booking package functionality to the TAFE Queensland Videolinq network. The refurbishment was researched, installed and implemented in a two-week period. The enhancements have enabled significant improvements in both the performance and availability of the network booking services. . Details on the performance improvements are outlined in the Annual Videolinq Report. An extract from that report is included below. TBA Web Server Upgrade A component telecommunications upgrade was finalised in December 2003 boosting bandwidth from 128 Kb availability to a synchronous 512:512 Kb configuration. Despite assumed access improvements as supported by a user survey following implementation, anticipated benefits failed to materialise when the booking system was put under load with the influx of new year bookings. Continued service difficulties prompted investigation of both the Qld TAFE data network and the existing server hardware. As a result of the analysis, an additional upgrade path was identified. Integrated Vision purchased a new Dell server to house the TBA Web page and its associated programs and provide the necessary data warehousing. This server runs dual versions of the TBA platform on ...

Informations

• Publié par: Enne
• Nombre de lectures: 25
• Langue: English

Extrait

TBA Web Server Installation & Performance Testing

TAFE QUEENSLAND Videolin TBA Web Server Installation & Performance Testing During February 2004, Integrated Vision undertook a major upgrade of the hardware infrastructure of the ‘To Be Announced’ (TBA) web server that provides conference booking package functionality to the TAFE Queensland Videolinq network.The refurbishment was researched, installed and implemented in a twoweek period. The enhancements have enabled significant improvements in both the performance and availability of the network booking services. . Details on the performance improvements are outlined in the Annual Videolinq Report. An extract from that report is included below. TBA Web Server Upgrade A component telecommunications upgrade was finalised in December 2003 boosting bandwidth from 128 Kb availability to a synchronous 512:512 Kb configuration. Despite assumed access improvements as supported by a user survey following implementation, anticipated benefits failed to materialise when the booking system was put under load with the influx of new year bookings. Continued service difficulties prompted investigation of both the Qld TAFE data network and the existing server hardware.As a result of the analysis, an additional upgrade path was identified.Integrated Vision purchased a new Dell server to house the TBA Web page and its associated programs and provide the necessary data warehousing. Thisserver runs dual versions of the TBA platform on individual hard drives to allow for redundancy processes. Additionally, server RAM was increased from 256Mb to 1Gb.Significantly, the entire TBA Web database now resides inside the available server RAM enabling immediate access to total data recall.This capacity dramatically improves the functionality and timeliness of the capability of the booking system. The new hardware configuration server has been commissioned and is operational.

3000 2500 2000 1500 1000 500 0

RAM (Mg)

Performance Testing

Report compiled  April 8, 2004 Revision 1.01 – May 6, 2004

TBA Web Server Upgrade

Processor Speed

IBM Server (old) Dell Server (new )

Author  Peter Anders Revision – Paul Crosisca

1

TBA Web Server Installation & Performance Testing

Following installation, internal service testing was conducted across randomly selected TAFE QLD network sites. Initial results identified positive outcomes. In order to benchmark performance formalised testing was conducted over a twoday period using a detailed survey that was sent to 20 sites across the TAFE network.A subsequent ‘load test’ was arranged for midafternoon of the second day with approximately 20 participants.The coordinated load tests provide the most comprehensive results supporting the improved functionality of the upgraded booking system capability. Participant List ‘Load test’ participants No. InstituteSites Personnel 1 CIDVideolinq OfficeShannon Jones 2/3 CentralQld MackayVal Farrell, Melanie Crawford 4 GladstoneNaomi Mason 5 GoldCoast SouthportCarolyn Tudehope 6 TropicalNorth CairnsHailee Russell 7 InnisfailKathy Hemming 8 BarrierReef InghamWendy Bonfield 9 PimlicoCheryl Harrison 10 SouthernQld RomaBeth Harms 11 KingaroyFay Groom 12 WideBay MaryboroughTrudi Tierney 13 HerveyBay JackiiHarrington Grey 14/15 BrisbaneNorth GatewayNatalie Small 16/17 OpenLearning OLIJanelle Devery, Jill Shelper 18 IVCSC MelbourneMiriam Krohn, Simon Adams 19 IVBrisbane PeterAnders Survey Contents Three different testing criteria and associated tasks were created to determine performance benchmarking across varying difficulty levels: ·Criteria 1 was tasked at minimum difficulty; ·Criteria 2  medium ·Criteria 3  maximum difficulty Criteria 1 a) Timetaken toconnect to TBA web via Videolinqwebsite and gain full screen functionality? b) Timetaken toconnect to TBA web via TBA web direct linkat 210.0.71.97 to gain a full screen? Outcomes 1

Report compiled  April 8, 2004 Revision 1.01 – May 6, 2004

Author  Peter Anders Revision – Paul Crosisca

2

TBA Web Server Installation & Performance Testing

a) Timeframes for access to full screen TBA functionality via Videolinq site ranged from 1 to 40 seconds giving a benchmarked average time across 19 sites of6.73 secondsshould be noted that when the excessive. It Maryborough time of 40 seconds is adjusted to a 10 second outcome, the average access time drops to 5.15 seconds. b) Benchmarkresults for full screen TBA functionality via direct link to TBA Web produces a4.63 secondaverage. Criteria 2 Checkroom availabilityacross 5 identified sites; Maryborough, Mackay, Cairns, OLI and Mount ISA  at an identified date/time; Friday 20 April between 10.00am and 12.00pm Outcomes 2 Time frames to complete this medium difficulty task of identifying room availability across 5 sites were between 2  40 seconds. The benchmarked average access time was7.52 seconds. Criteria 3 Place and receipta standard conference booking of 1 hour duration at unspecified day / time across 5 network sites Outcome 3 This task provided the most unified results with the average booking time across all sites between 1 and 3 minutes.The formal benchmark across 19 sites for the completion of thishigh difficulty taskwas1 min 52.63 secs Tabulated results of the BENCHMARK testing appear below and on the next page. Benchmarked Average Outcomes

TBA WEB Server: Performance Testing

Benchmarked Avera

120 110 100 90 80 70 60 50 40 30 20 6.73 10 0 Criteria Tests Videolinq Access

Report compiled  April 8, 2004 Revision 1.01 – May 6, 2004

4.63

Direct IP Access

7.52

Room Availability

112.63

Book & Receipt

Author  Peter Anders Revision – Paul Crosisca

3

TBA Web Server Installation & Performance Testing

TBA Web Performance Testing: Individual Site Data ParticipantsCriteria 2Criteria 3Criteria 1 (secs) secs1b (secs)1a (secs) CQIT Melanie Crawford (Mackay)2 22 120 Val Farrell (Mackay)1 13 180 Shannon Jones (Videolinq)5 52 120 Naomi Mason (Gladstone)12 68 60 TNQIT Hailee Russell (Cairns)3 53 60 Kathy Hemming (Innisfail)3 38 60 BRIT Wendy Bonfield (Ingham)5 57 180 Cheryl Harrison (Townsville)15 152 180 WBIT Jackii HarringtonGrey (Hervey Bay)2 11 60 Trudi Tierney (Maryborough)10 3 30* 180 SQIT Beth Harms (Roma) 40* 10 15* 30 Fay Groom (Kingaroy)7 5 40* 240 BNIT Natalie Small (Gateway)7 93 180 OLI Janelle Devery1 34 70 Jill Shelper1 34 60 GCIT Carolyn Tudehope (Southport)2 22 180 Integrated Vision Miriam Krohn (CSC Melbourne)1 11 60 Simon Adams (CSC Melbourne)7 53 60 Peter Anders (Brisbane)4 45 60 Total (secs)128 88143 2140 Benchmarked Average (secs)6.73 4.637.52 112.63 *‘Slower’ comarative access exerienced at the two SQIT sites and Marborou h

Report compiled  April 8, 2004 Revision 1.01 – May 6, 2004

Author  Peter Anders Revision – Paul Crosisca

4

TBA Web Server Installation & Performance Testing

Additional Comments CQIT Mackay– Val Farrell:“TBA is operating at a high standard and is a pleasure to use now” CQIT Gladstone– Naomi Mason: “TBA is much quicker than before and it’s displaying everything on thepage” SQIT Kingaroy– Fay Groom: “Screen changes were the quickest I have ever seen” WBIT Maryborough Trudi Tierney: “ Switching between screens has only taken 1 2 seconds. Very fast “ BRIT Townsville– Cheryl Harrison: “The performance was excellent overall. TBA is usually so slow“ TNQIT Innisfail– Kathy Hemming: “An efficient, user friendly system. Fast and quick to load“ BRIT Ingham– Wendy Bonfield: “Thank you for the opportunity to test the TBA. It was great and I can’t believe how fast it was. It was truly amazing.” BRIT Charters Towers– Robyn Smith: “This is working instantly today. I was hoping that today’s experiment would reflect the problems that I have had but I have rarely got this sort of prompt response before” Conclusion The overall performance and quality of service of Integrated Vision’s TBA web booking system has dramatically increased through the implementation of the new Dell PowerEdge server and the associated hardware upgrades. As documented by the test results over several days including the Load Test, the new server can maintain multiple users simultaneously accessing the TBA Web booking system.With a number of users logged in, the functionality of the new server is excellent compared to previous service capability. The average time for booking enquiries, processing and lodgment has decreased dramatically, as per the benchmarked averages and site response times detailed in this document by relevant tables and graphs. Overall, the fastest access to TBA is via the direct IP address however; this may be attributed to congestion factors affecting access to the Videolinq internet site, currently housed on a server at OLI. It should also be noted that access at the two SQIT campuses, whilst improved, is still ‘slow’ compared with other campuses. Further justification of dramatic improvement is drawn from anecdotal comments and informal testing results provided by network personnel across the state.The positive feedback provided by CQIT and SQIT campuses is particularly relevant given that these were previously experiencing the vast majority of TBA connectivity problems. A sincere thank you to all participants for assisting us with this testing process. End document

Report compiled  April 8, 2004 Revision 1.01 – May 6, 2004

Author  Peter Anders Revision – Paul Crosisca

5