William Faulkner once said, “The tools I need for my work are paper, tobacco, food, and a little whiskey.” For IT connectivity support, additional tools are required, not to say that some of Faulkner’s don’t also come in handy. Thanks to the work of folks in ITS Networking and our friends in the ITS Control Center there are a lot of useful tools available to campus support folks that everyone should keep in mind when analyzing connectivity issues.
Let’s start with the toolkit that’s available at the ITS Control Center. If you look at the Links section on the right side of the page, under Applications, the places to always start for any connectivity issues are going to be the status.its page (aka Message Center) and Cujo. The status.its page contains information about any scheduled or unscheduled but noted activity or outage that’s taking place. If you click on the Message associated with any event, you’ll get more details about that item.
Another essential ITS Control Center tool is the Cujo application. Cujo pings thousands of network switches, wireless access points and critical hosts approximately once every 90 seconds and displays any outages that it discovers. This should be your first point of contact with any reports of loss of connectivity. For a single user unable to connect to the network, the Network Information Tool (https://nit.unc.edu) should be checked to determine if a given MAC address has been “blocked” or put into a “Penalty Box” state. This tool will not only indicate if a MAC address has been blocked from the network, but what Remedy ticket number exists for this incident.
The Network Information Tool can provide you with more information that just whether or not a device on the network is blocked. You can find out where the device is connected, at what speed it’s connected and what VLAN it’s in. For more information about this tool, see http://help.unc.edu/help/network-user-information-tool/
Other than basic connectivity issues, another area where tools come in to play is in terms of performance measurement. If you feel you’re getting poor network throughput, is it a network issue, an application issue, or the dreaded “something else”? The tool to start with in these situations is the Network Diagnostic Tool (or NDT). This tool uses use a Java applet to run a bidirectional network throughput test of 10 seconds in each direction between your computer and that NDT system. The link for the primary NDT server deep within the ITS data center VLANs is http://ndt.itcc.unc.edu:7123; this would most closely approximate your throughput to main ITS central services. There are also NDT servers within the School of Medicine at http://ndt.med.unc.edu:7123 and the School of Business at http://bschool-ndt.net.unc.edu:7123. In the coming year, we’re looking to upgrade the connectivity to the NDT servers to 10Gbps. This definitely is a case of “your mileage may vary” – you may have a gigabit interface on your computer, but are connected to a 100 Mb/sec switch port; or your computer may have a large file transfer going on in the background, so you should consider that when you run this test and you should run it multiple times to get an average baseline. However, if you get throughput measures of 500-800 Mb/sec on a gigabit connection, that’s generally good throughput; if you get measures of 60-80 Mb/sec on a 100 Mb/sec connection, that’s likely as good as you’re going to get from a network perspective. If you get single digit throughput or worse in Kb/sec, then there’s something going on that needs further investigation from a network perspective. Note that if you have a system that cannot run Java applets (like any tablets on the market today), you would not be able to use this tool.
Also related to bandwidth indicator tools is the Cacti page, found from the Control Center page or directly at http://cacti.unc.edu/cacti – note that this page requires Onyen authentication to access. Once you access that site, you’ll want to click on the Weathermap tab at the top of the page. The Weathermap site visually depicts the basic connectivity architecture of the campus network in five different views: (1) network core (2) main campus (3) ITS data centers (4) ResNET (5) Med School. Each inter-switch/router link can be clicked to see current and historical bandwidth and most switches and routers can be clicked to see individual interfaces on those devices.
We often get tickets about poor wireless performance in areas that we may not have targeted wireless coverage. An important place to start in resolving wireless issues is verifying that you have an access point in an area. Use the campus wireless access map, to find links to campus maps indicating for every building on campus that has one or more WAPs exactly where those access points are installed. Given the vagaries of wireless clients, the exact level of signal strength you’d get would vary, but you can probably be sure that if you’re on the 6th floor of a six story building and the only APs are on the 1st and 2nd floor, you’re not going to have good connectivity.
These are only some of the tools that we’ve been able to make available for basic information gathering and to use as a first stop in troubleshooting. Please feel free to check out these tools and determine which ones are best for your situations. If there are other types of connectivity analysis tools that you would find useful, please reply directly to me and we’ll see what our teams can come up with.
Jim Gogan – Information Technology Services (ITS)