A NEW PARADIGM FOR
NETWORK PROBLEM SOLVING
Are Problems a Thing of the Past?Every year, networks become more reliable. New standards promise interoperability, new devices simplify configuration and advanced monitoring solutions offer to detect problems before users are impacted. IT departments are entering an era where problems will be a thing of the past.
Or are they? A recent research study of over 300 network professionals in large - and medium - sized organizations found that:
- 48 percent of all organizations average longer than half a day to close trouble tickets
- 46 percent of all organizations are under pressure to reduce the time it takes to close trouble tickets
- Network professionals spend about 25 percent of their time solving problems
To increase productivity, troubleshooting tools not only need to keep up with technology changes, but must continue to improve processes used to solve problems.
How Troubleshooting is Done TodayThe vast majority (72 percent) of organizations do not follow a standardized troubleshooting process. Not only does this process vary within an organization but the tools used to troubleshoot problems vary substantially. Survey respondents reported using eight different types of tools to solve problems. In 47 percent of the situations, two or more tools were needed. With all the variability in troubleshooting practices and tools, it’s not surprising that 63 percent of troubleshooting sessions lasted more than an hour.
There is a part of problem solving that is worth considering separately. In many cases, technicians can’t resolve the problem themselves. Sometimes they need additional help with especially difficult problems. In other instances, it’s because the problem lies outside their domain of responsibility, and they need to work with a separate group inside (server management or application developers) or outside (service providers or equipment vendors) the enterprise. This is far from a rarity – our research indicates that 41 percent of all issues require collaboration of this sort. This can take too long for at least two reasons. First, it’s not always easy to give the responsible parties visibility to the problem when it’s occurring. Second, the technician may not have the ability to easily capture the trace files that are often required (19 percent of the time) for these problems.
A Study of Problem-solving TechniquesThis white paper refers to a NETSCOUT research study of 315 network professionals in April 2012. The respondents came primarily from medium- to large-sized networks in a variety of industries. Most of them were top-level networking support staff.
Figure 1: Which of these tools did you use to troubleshoot your most recent user problem?
Figure 2: How many trouble tickets do you process in a typical month?
Figure 3: What is your group’s average time to close a trouble ticket? Note that 48% report more than four hours.
Figure 4: What was to root cause of the last user problem you solved? (Multiple responses allowed.)
Changing the Problem-solving ParadigmAt NETSCOUT, we looked to shorten the entire problem-solving process. The process, as described earlier, traditionally consists of two steps – solo troubleshooting and collaboration when necessary. To streamline troubleshooting, NETSCOUT developed a three-step process and designed a new tool based on it. The three steps are:
1. Automated Testing
The OneTouch™ AT Network Assistant, enhances each of the steps and greatly reduces the time to solve problems.
Step 1: Automated TestingIt may seem counterintuitive that adding a step reduces time. But if the additional step actually saves more time in subsequent steps, the total time is reduced. That’s the idea behind automated testing.
The OneTouch AT identifies the most common network problems in about one minute1. It performs a thorough network analysis from the end-user’s point-of-view. Such an analysis, performed manually, would take roughly one hour2. The OneTouch AT performs that same testing in about a minute (tests are configurable and can take anywhere from ten seconds to a few minutes, with most being less than a minute). Results are then compared against user-defined limits to provide a simple pass/fail result. This approach allows technicians to find the most common problems that result in end-user complaints.
There are several benefits to automated testing. First, it’s much faster than a typical trial-and-error test. Second, it’s more thorough than a manual approach, which means it can find problems that the technician may not have even considered. Third, it allows anyone, regardless of skill level, to run tests and identify problems.
Step 2: TroubleshootingWhile the AutoTest uncovers a wide variety of problems on its own, not all problems can be found that way. The OneTouch AT provides a veritable arsenal of troubleshooting power to reduce the time spent in this phase.
AutoTest – even if the initial AutoTest does not identify the problem, all the measurement results are ready and available to help the technician understand what is happening. Further, the AutoTest can be modified in seconds and re-run to test a different server, application or wireless connection.
Wired Tests – a complete set of tests provide information on the cable, Power over Ethernet, the nearest switch and network services. The OneTouch AT features a web browser, Telnet and a SSH client to assist with configuration of network devices including switches and access points. It features a toner and can flash switch port lights to help locate unmarked cables in congested closets. A video probe can be connected to the OneTouch AT to inspect the endface of fiber optic connectors for contamination.
Step 3: CollaborationAs noted earlier, network technicians regularly need to work with someone else to resolve problems. The process of getting the right information to the right people, however, can drag on for days. Even if the technician is able to work on other problems during this period, that’s little comfort to an end user who can’t get their job done or the IT manager missing targets for trouble ticket times.
The OneTouch AT includes features specifically designed to expedite collaborative troubleshooting with the Link-Live Cloud Service. With Link-Live Cloud Service, everything is always captured with zero-touch reporting, providing you with complete records of your work and results, at all times. This comprehensive, searchable database allows you to easily identify changes in the network. There is no space limit, no time limit, and no need to retrieve and transfer data from testing devices to storage at the end of the day. This cloud service is available to all Handheld Network Tools: LinkSprinter, LinkRunner AT, and the AirCheck G2, all automatically upload their test results at the moment they happen. If you don’t have connectivity at that drop, the testers can buffer results, which are pushed to the Link-Live Cloud Service when connectivity returns. Technicians no longer have to be the keepers and transporters of test data, or even remember to press save. Instead, they can double their productivity and finish jobs that used to take weeks, in days. In addition, managers can have access to the test results dashboards.
Reporting – a detailed report of everything that the OneTouch AT tested and observed will be recorded in the Link-Live Cloud Service. This allows the tech to show a colleague exactly what is happening when they are observing the problem. These results include results a less-experienced technician might not have looked at but are there for more knowledgeable team members to evaluate.
In-line Packet Capture – a trace file is indispensable for very difficult problems or as evidence to an outside group such as application developers, service providers or equipment suppliers. Collecting this information typically requires reconfiguration of the switch or a network tap. This can take 30 minutes or more. Worse, many techs may not have access to switch provisioning or a tap. That means even more delay as the problem is escalated to another individual.
The OneTouch AT can perform an inline packet capture in just a few screen touches without the need to access the switch or a tap. This means the tech can capture the problem packets immediately while the user demonstrates the problem.
Web Remote Interface – while it’s not always possible to get a colleague physically in the location of the problem, the OneTouch AT can be accessed and controlled through a remote device such as a PC, tablet or smartphone. Not only can the remote user see what the tech is seeing, but they can control the OneTouch AT and export trace files or reports to their device.
SavingsThe first step in estimating savings from the OneTouch AT is to look at the time saved in each of the three parts of the test.
Automated Testing – Table 1 compares the amount of time it would take to perform the AutoTest functions with the actual time of the AutoTest. The time is dependent on the skill of the technician as well as how many applications would need to be tested.
Troubleshooting – it’s less straightforward to quantify these savings, as it is highly dependent on the actual problem and the skill of the technician. Users of other NETSCOUT Handheld Network Test Solutions generally report 30 percent to 40 percent faster troubleshooting, but we will set that aside and consider it a “bonus” savings in addition to what is demonstrated here.
Collaboration – to quantify the time savings in these scenarios, we compared time to set up a packet capture using port mirroring (roughly 20 minutes) versus the inline packet capture with the OneTouch AT (three minutes). In instances where the technician does not have access to reconfigure the switch, the time savings would be much greater.
Better collaboration offers a bigger benefit by reducing the overall time to close trouble tickets. Without the OneTouch AT, it’s often difficult to get all the relevant data in front of the appropriate people quickly. This is a significant cause of trouble tickets dragging on, sometimes for days. Armed with the OneTouch AT, the first responder easily generates a report or a trace file and shares these plus provides remote access for the rest of the team in real time. So while the total amount of time the staff spends on a problem may not be reduced, the time the end user spends waiting for a resolution is greatly reduced.
Again, such a benefit is hard to quantify – so we won’t – but for many organizations this may be more valuable than actual hours saved for the department staff.
ConclusionsAn estimation of the savings expected from using the OneTouch AT is presented in Table 1. Even ignoring the time saved in solo troubleshooting and purchase of the top of the line model, one would expect payback in less than six months.
|Trouble tickets per month per tech||20 (median)|
|Minutes per ticket||90|
|AutoTest time||1 minute|
|Time to manually perform autotest functions
(from Table 1)
|Time saved per ticket||59 minutes|
|Hours saved per month||19.7 per tech|
|Percentage requiring packet capture||19% (average)|
|Number requiring packet capture||3.8|
|Time for packet capture setup||20 minutes|
|Packet capture setup with OneTouch AT||3 minutes|
|Time saved per capture||17 minutes|
|Hours saved per month||1.1 per tech|
|Number of users||2 per OneTouch AT|
|Total time saved||41 hours per month|
|Total monthly savings||$2,489|
|OneTouch AT cost||$10,000|
Table 1: An estimation of time and cost savings provided by OneTouch AT.
Figure 5: Summarizes the concepts of this paper: the ways in which the OneTouch AT saves time throughout the entire problem-solving process as compared to traditional methods.
Appendix A – An Hour of Troubleshooting in One MinuteGetting to the root cause is critical but can involve extra steps. For example, when you go to the doctor’s office, regardless of your complaint, you’re first seen by a nurse who immediately measures your weight, temperature, blood pressure and often reviews your health history. This step not only saves the doctor’s time, but often catches issues that might otherwise be overlooked.
This concept also applies to testing a network. Since your team doesn’t usually include a nurse, the OneTouch AT automates a complete test of “network vital signs” into an AutoTest that compresses an hour of traditional testing into about a minute. The results are then compared against userdefined limits to provide a simple pass/fail result. This approach not only saves time, but allows technicians to solve more problems.
|Testing Step||OneTouch Network
|Basic Connectivity (Wired or Wi-Fi)||1 minute||Cable Tester, PC, Wi-Fi Utilities||5 minutes|
|Infrastructure Services||PC, Utilities||5 minutes|
|Wireless Operation and Performance||Two PC’s, iPerf||10 minutes|
|Network Services and Application Performance||Packet Capture, Protocol Analyzer||40 minutes
Table 2: The OneTouch AutoTest performs nearly an hour of manual testing in one minute.
Infrastructure Services – OneTouch AT tests availability and response time of DNS and DHCP across both the wired and wireless network.
Wireless Operation and Performance – the exclusive test measures the actual wireless performance by sending a stream of traffic out the wireless port, through the nearest AP, the wired infrastructure and back to its wired port. The test runs simultaneously in the reverse direction with programmable upstream and downstream rates. The test provides measurement results for throughput, loss, latency and jitter in both directions. These latter two measurements are vital for quality performance of real time application such as streaming video or voice over Wi-Fi.
Network Services and Application Performance – these tests provide a detailed breakdown of application performance by analyzing an actual interaction with the server/service under test. Seven different tests/applications are supported and can be customized for specific sites or applications whether hosted locally or through cloud-based providers.
The detailed breakdown provides the results shown in Figure 7 and shows the results for both the wired and Wi-Fi network side-by-side for easy comparison.
While the overall measurement can be used to quantify the performance of the application, the detail can be used to determine the cause of the slow performance, whether it’s the network, application server, or the DNS server. While an expert can provide this level of analysis in ten or twenty minutes with a protocol analyzer (once they have a trace file), the OneTouch AT provides it under a minute as part of a standard AutoTest.
Figure 6: The test measures performance of the wired and Wi-Fi network.
Figure 7: The application performance tests provide a detailed breakdown of the response time of servers and services.
Appendix B: List of problems that can be discovered by the OneTouch AT Network Assistant AutoTest
|Fiber problems||Wrong SFP/vendor mismatch, Dirty fiber endface*, Dead port/broken fiber, Low power|
|Twisted pair problems||Open cable, Bad cable mapping, Shorted cables, Mislabeled/undocumented cables*, Too long cable, Dead port|
|PoE||Not present or disabled, Switch unable to supply adequate power, Wrong pins, Low voltage, Non-Ethernet voltage, Low power under load, Class 4 negotiation mismatch|
|Link||Polarity mismatch, Low link level, Receive pair issues (MDIX), Speed mismatch, Duplex mismatch|
|Switch port||Incorrect switch, Incorrect port, Incorrect data VLAN*, Incorrect voice VLAN*, Unstable switch uptime*, Switch congestion*, Switch errors*, FCS errors*, Frame size errors*, Other frame errors (7)*, Excessive broadcast traffic*, Excessive multicast traffic|
|Wi-Fi||Security settings wrong, AP missing, AP misconfigured, AP not connected, WLAN controller problems*, Excessive noise*, AP congestion*, Channel over utilized*, Too many APs on channel*, AP overlap on channels*, Roaming problems*, Unauthorized APs*, Finding rogue APs*, Ad hoc networks*, Insufficient network coverage, Slow connection, Bandwidth hogs APs and client*, Bad client NIC*|
|Veri-Fi||QoS settings wrong*, MTU problems*, Port problems*, Upstream bandwidth issues, Downstream bandwidth issues, IPv6 issues*, Excessive loss, Latency issues*, Jitter issues*, Sequencing issues*|
|DHCP||Missing, Slow, Out of addresses, Incorrect lease time, Rogue DHCP server, Wired versus Wi-Fi configuration problems, Duplicate static IP address, IP address hijacked, Incorrect address delivery, Incorrect subnet delivery, Incorrect router address, Incorrect DNS address|
|DNS||Missing, Slow, No secondary server*, Wired versus Wi-Fi configuration problems*|
|Gateway||Missing / Failed, Not IPv6 capable*, Unstable gateway uptime*, Overloaded*, Bad traffic*, Incorrect routing protocols*|
|Discovery||Wrong VLAN*, Wrong subnet*, Unexpected IPv4/IPv6 devices|
|Web (HTTP)||DNS lookup failure, DNS lookup slow, Server unavailable, Slow connectivity, Server slow to start, Server slow to complete, Wired versus Wi-Fi transport problems, IPv4 versus IPv6 transport problems|
|Ping (ICMP)||DNS lookup failure, DNS lookup slow, Server unavailable, Slow connectivity, MTU misconfigurations, Wired versus Wi-Fi transport problems, IPv4 versus IPv6 transport problems|
Appendix B: (continued)
|Connect (TCP)||DNS lookup failure, DNS lookup slow, Server unavailable, Slow connectivity, Firewall misconfigured for ports, Wired versus Wi-Fi transport problems, IPv4 versus IPv6 transport problems|
|Multicast (IGMP)||Server not multicasting, Switch IGMP snooping disabled, Incorrect port configuration, Server authentication|
|File (FTP)||Slow WAN, DNS lookup failure, DNS lookup slow, Server unavailable, Slow connectivity, Server slow to start, Server slow to complete, Wired versus Wi-Fi transport problems, IPv4 versus IPv6 transport problems|
|Video (RTSP)||Slow WAN, DNS lookup failure, DNS lookup slow, Server unavailable, Slow connectivity, Server slow to start, Server slow to complete, Wired versus Wi-Fi transport problems, IPv4 versus IPv6 transport problems|
* May require a secondary test after the AutoTest.
1 See appendix A, “An Hour of Troubleshooting in One Minute”
2 See appendix B, “Problems That can be Discovered by the OneTouch AT Network Assistant AutoTest”