• NCDesk
  • NCDesk Home
  • Download
  • Versions
    • About Versions
    • Download Current Release
    • Release History
    • Autoupdate
    • Demonstration Apps
  • Installation Instructions
  • Autoupdate
  • Helpdesk
  • Error Codes
  • Bug Reporting
  • Sample Files
  • File Conversion Guide
  • Test Simulation
    • About the Test Simulation
    • Test Simulation Demo Movie
  • Features Release Schedule
    • NCDesk (Global)
    • Editor
    • Database
    • Spreadheet
    • Window Manager
    • Mail
    • Multimedia
  • Technical Notes
    • Technical Notes Home
    • Spell Check Dictionaries
    • System Architecture
    • Autoupdate
    • Bandwidth-Network Use
    • Bandwidth-Footprint
    • Host Network Analysis
    • Testing Server
    • Requirements
      • Requirements Overview
      • Proposed Network Requirements
      • Proposed Client Requirements
    • Special Considerations
  • Test Registration System
  • Required Technical Project Management Documents
• New Test
  • New 2004 Test Home
  • Background
  • Strand Percentages
  • Events Planned for 2007-2008 School Year
  • Alternative Assessment
  • Accommodations
    • Approved Accommodations
    • Computer Reads Test Aloud
    • JAWS
  • Official actions-recommendations
  • Research/Studies/Results
    • Fall 2003 Study
    • Fall 2004 Trial
    • Spring 2005 Field Test
  • Goals
  • FAQ
    • All FAQs
    • Test Design
    • Accommodations/Special Needs
    • Test Simulation
  • Irregularity Reporting
  • Presentations
  • Guides & Manuals
    • Test Simulation Demo Movie
    • Administration Information Feasibility Study (Fall 2004)
    • Field Test Administration Manual (Spring 2005)
    • Administrative Information for Alternate (Spring 2005)
    • Test Administration Manual (September 2005)
  • Physical Test Environment
  • Test Security
  • Related Links
• New 2004 SCS
  • NCDPI Computer/Technology Skills SCS
• Other Tests Links
  • NCDPI Computer Skills Test Web Page
  • 1998 Curriculum Test (Archived Information)
    • 1998 Curriculum Test Handbook Website
    • 1998 Curriculum Test Handbook PDF
    • 1998 Curriculum Test Performance Sample Items And Files
  • 1992 Curriculum Test (Archived Site)

NCDesk

Technical Notes - Host Network Analysis

During the spring of 2005 field test an analysis of the network supporting to host systems was conducted by the Centennial Networking Lab. The analysis supports a conculsion that available bandwidth for the online testing services is more than adequate for handling the load of this test with additional room for growth.

The report was based upon the statement of work written by Centennial Networking Lab (CNL) for NCSU Center for Urban Affairs & Community Services (CUACS). The test plan was as follows:

• Determine path(s) to internet/NCREN from testing host server room
• Determine end-point(s) for testing based on information above
• Characterize the path from testing host server room to end-point(s)
• Validate NCDesk bandwidth requirements (if desired)
• Monitor bandwidth utilization at testing host server room, building exit, and points beyond as available

Campus Network Diagram

Testing Host Services Access

The testing host services connect via 2Gbps (1Gbps full-duplex) fiber to a router in SMDF. SMDF connects to the campus backbone via a 4Gbps (2 x 1Gbps at full-duplex) fiber link. The campus core consists of two layer-2 switches connected at 8Gbps (4 x 1Gbps at full-duplex).

NCSU bandwidth to UNC, Duke and NCREN is 1Gbps (upgrading to 10Gbps soon). From NCREN to other sites, the bandwidth varies according to traffic management policies; minimum 100Mbps to a maximum of 2.48Gbps for Internet2 connected institutions. [DPI connectivity is 45Mbps thru ITS to NCREN (last time we tested it).]

Testing Host Traffic Utilization History

The following historical traffic data was collected May 4, 2005.

Network Traffic Generation and Collection

For the period: May 24, 2005 to June 29, 2005, the following data was collected using the iperf performance tool.

Machine: data7.ncsu.edu
Minimum Throughput: 80.8 Mbits/sec
Maximum Throughput: 85.7 Mbits/sec
Average Throughput: 85.4 Mbits/sec
Standard Deviation: 1.03
Samples: 1761

The following graph shows the data collection stopping in the middle of week 26.

Analysis and Results

From the data above, the building hosting testing services has very low traffic utilization average (4.5Mbps, 6Mbps during collection period) and has ample available throughput thru the NCSU core (minimum 85.7Mbps). The 85.7Mbps measurement represents a hardware limitation of the machine used as the end-point for the testing.

The lower throughput measurements (80.8 to 84.9Mbps) were always seen between 6am and 7am. This probably coincides with some local data backup by some service in the building telecom closet; the last graph above shows data transfer occurring between 5am and 7am with sustained rates between 14Mbps to 18Mbps for approximately an hour and a half.

The average utilization of the campus backbone is on the order of 300Mbps.

In summary, NCSU provides ample bandwidth for network related services. When creating services for customers over the Internet, the likely bottleneck will be at the last-mile to the school/user.

This file was modified on Wednesday, August 29, 2007; at 1:36:41 PM