Campus Wi-Fi infrastructure and quality assurance

The University of Waterloo operates one of the largest campus Wi-Fi services in Canada, with over 5,500 Wi-Fi access points (APs) serving over 30,500 people per day.

We do everything we can to ensure a good user experience, from commitment in staff time and budget to campus Wi-Fi infrastructure and service, use of leading technology, a validated design approach,  infrastructure monitoring, and ongoing quality assurance.

The service you receive is dependent on many factors, including the equipment you use, and how it is configured.  For more information on the factors that affect Wi-Fi service, and what you can do to receive the optimum experience, please read Factors affecting Wi-Fi service.   If you are having difficulty with the University’s Wi-Fi service, please see Connecting the the eduroam wireless network.

For more information about the University’s role in providing campus Wi-Fi service, please read on...

Upgrades

  • Upgrades of entire campus Wi-Fi infrastructure to dual band a/b/g/n completed (excluding Affiliated and Federated Institutions of Waterloo).

  • Campus Wi-Fi upgrade (now complete)

    • Waterloo Daily Bulletin - March 15, 2013

    • Information Systems & Technology (IST) is in the final year of a three-year, $1.2M upgrade to wi-fi in all academic and administrative buildings across campus…

  • Residence Wi-Fi upgrade (now complete)

Infrastructure

The University uses leading vendors for network infrastructure:

  • Aruba Networks for Wi-Fi

  • Cisco for core routing and NAT infrastructure

  • Cisco for access layer wired infrastructure

  • Sandvine for traffic management

  • Gartner Magic Quadrant for the Wired and Wireless LAN Access Infrastructure Published: 13 June 2012

Campus Wi-Fi design

  • The design of Wi-Fi for a new building, or re-design of Wi-Fi in an existing building, starts with loading floor plans and information about building materials in Airmagnet Survey Professional.

  • Including capacity information about lecture rooms and meeting rooms.

  • Pre design walk through of building (for existing buildings).

Airmagnet survey professional

  • Place access points onto the floor plans (predictive modeling) to meet the following conditions:

    • Access points in classrooms to provide approximate capacity of 50% of seats using wifi and maximum 25 devices per radio: limitation being there are only 3 2.4Ghz channels.

    • minimize co-channel interference by adjusting predictive model (while maintaining coverage).  Co-channel interference is caused when access points on the same channel pick up each others signal, and is a particular problem on the 2.4Ghz band.

    • predictive modeling is done to meet a signal requirement of -79dbm on the DWA-160 adapter, (-70dbm on an average adapter).

Airwave management platform

  • Airwave contains floor plans and AP locations for entire campus

  • Provides coverage heatmaps, usage history, userids, noise and throughput history on each AP radio

  • And much more

Post install survey to confirm coverage and co-channel interference

  • Post install survey with Airmagnet is done in the public areas that are accessible (hallways near offices, classrooms if not in use, labs if allowed to enter)

    • Includes coverage and co-channel interference on both 2.4GHz and 5Ghz band

  • Adjustment to AP power levels if necessary to reduce co-channel interference or increase coverage

Diagram on right is 2.4GHz coverage. Diagram on left is co-channel interference on 2.4GHz. There is typically no co-channel interference on 5GHz band.

Ongoing quality assurance

  • The use of leading network technology, a validated design approach, commitment in staff time and priority in budget,  assures a robust campus Wi-Fi infrastructure.

  • We also monitor the infrastructure, monitor some elements of user experience (where possible), conduct random spot tests, conduct satisfaction surveys, and track usage of all elements of the Wi-Fi infrastructure.

  • All of the IST staff that manage UWaterloo’s Wi-Fi infrastructure are regular users of the service.

  • We investigate user reported issues, and can enhance coverage if determined to be an issue.

  • Unfortunately…

    • Interference from non Wi-Fi sources, and rogue APs can impact service.

    • Issues on client devices (driver bugs, poor implementations, etc) can impact end user experience.

    • There is a limit to the number of 2.4GHz APs that can be enabled in large lecture rooms or other large open areas like the DC library, due to co-channel interference, and only 3 channels available in the 2.4GHz band.

Infrastructure monitoring

  • Monitoring health of core controllers, routers, and traffic management, with automated alerting.

  • Detecting and replacing failed APs, through automated alerting.

  • Detection of heavily used APs and locations, and addition of APs if needed (if possible, given limitations of 2.4GHz spectrum).
     

Spot tests

  • IST co-op student visits areas of campus and tests Wi-Fi service.

  • Approximately 6-10 hours per week.

  • Approximately 20-30 areas visited per week (typically higher student load areas).

  • Combination of subjective tests, and non subjective tests.

  • Results recorded in a database with locally developed spot test tool and system.

Spot test results

(This highlights the challenge of a low quality adapter (DWA-160) in high usage areas on campus.)

Band

Adapter

Tests

Passed

Success %

Band

Adapter

Tests

Passed

Success %

5Ghz

Intel(R) Centrino(R) Ultimate-N 6300 AGN

156

156

100

5Ghz

D-Link DWA-182 Wireless AC Dual Band USB Adapter

26

26

100

5Ghz

ASUS USB-AC56 802.11ac Wireless USB Adapter

25

25

100

2.4Ghz

Intel(R) Centrino(R) Ultimate-N 6300 AGN

103

99

96.1

2.4Ghz

Intel(R) Centrino(R) Wireless-N 1000

81

81

100

2.4Ghz

D-Link DWA-160 Xtreme N Dual Band USB Adapter(rev.A2)

45

19

42.2

2.4Ghz

ASUS USB-AC56 802.11ac Wireless USB Adapter

3

3

100

Quality of experience

  • Sandvine traffic management system also provides video quality of experience through its ‘Real Time Entertainment Dashboard’.

  • This provides an estimate of the users quality of experience by examining traffic behaviour through network.

  • Below shows YouTube quality on campus Wi-Fi (out of five).

  • Improvement aligns with campus Wi-Fi upgrade.

Client evaluation of service (at the point of service)

  • We randomly invite approximately 60 active Wi-Fi users per day to visit a web page and provide feedback

  • Campus and Residence results are maintained separately

    • Sample email:

      Your input on the University of Waterloo's Residence Wi-Fi service is requested

      We'd like you to visit:

       https://istns.uwaterloo.ca/wifiservicereport?key=58032dda3c469893376d76e4c523599

      and answer a few questions, and submit the results. No userid or password is required.

      Your participation is voluntary.  You will not be sent any followup email, or another invitation to participate, within two years.

      If you'd like to participate, please complete the survey before Oct 09, 2013 23:59:59

      Individual responses will be kept confidential, and destroyed after two years. Summary data which excludes individual responses may be published.

      Please do not reply to this email

  • Example Survey Screen:

Survey results

  • About 15% of users respond to the Wi-Fi survey

  • Can view total responses in pie chart form

  • Results for Housing residences shown…