What is a broadcast storm?

Under certain conditions, incorrectly configured routers/bridges can pass a broadcast between segments in a ring - forever.

Symptoms

The network becomes clogged with unnecessary packets, response times increase dramatically.

How to find it

In the TRAFFIC MONITOR find the broadcast address (usually called Broadcast, it is an address of FFFFFFFFFFFF).

Click the right mouse button on the Broadcast node, select Conversations, Node to Node, and step through the list to find the conversation with the exceptionally high number of broadcasts. This isolates the node and therefore the segment containing the problem.

Refer to the router information to rectify the problem.

Analyzing Remote Segments

Often a central office has a requirement to monitor, or perform health checks on remote sites. The Installation has two parts, one involves loading the Podlink on the remote segment and the second involves configuring the management station to communicate with the Podlink.

Setup a Podlink on a Node on the remote segment

To monitor a segment you need to setup a Podlink in a machine on the segment. Only one copy of Podlink is required for each segment.

• Insert the CD in the machine where you want to run Podlink.
• Select Browse this CD.
• Select Podlink Installation.
• Save the file SETUPEX.EXE on to your local hard disk.
• Run SETUPEX.EXE and follow the on-screen instructions.

Connecting to a remote segment

The Distributed Analyzer can be used to monitor remote segments.

Add a new segment to your Segment Manager screen.

• Click on the New Seg button.
• Type a name for the segment, and select a bitmap to represent the segment.
• Click the Link to Pod tab.
• In the Podlink to Podlink section click on the drop down list next to Internal. Click on Remote over TCP.
• Type the IP address (or DNS name) of the remote node running the Podlink software.
• Click OK.

If you need to check the settings on a segment icon click the right mouse button, Properties, Link to Pod tab.

Why have a Conversation Table?

It is obviously important to know who is on the LAN at any given time and how busy the network is. However, it is not enough to leave it at that. You need to know more. Which nodes are the servers talking to? When did the server last send a packet to an individual node? when (if at all) did the node last reply?, which protocols are being used? (not just generally, but on a node by node basis), how many packets have been sent and received by each node ? A Conversation Table, keeping track of all the connections between nodes and servers, all the protocols, and all the traffic details, is the answer.

For example, you may have a node incorrectly configured to receive IP traffic and a server sending Novell IPX packets. A look at the source address (the server) would show packets leaving and addressed to the node. The node would not accept the packets since they are not of the correct type. The node would try sending packets to the server announcing its existence, but the server would not accept the packets either. With a Conversation table it's obvious what is happening. The server and the node each show up as two conversations (one with IPX traffic, one with IP traffic) with traffic in only one direction. Without the protocol information it would be a much more difficult problem to spot.

How to find the Conversation Table

The Conversation Table is accessible from many places in the software.

For an Individual Node

• Click the right mouse button on a node, select Conversations.
• Highlight a node, click the Conversations button.

This also gives the color coded status of each Conversation, green for a two-way conversation without errors. Blue for a single-sided conversation, for example to Broadcast or Multicast addresses, red for a conversation containing errors, and dotted for a conversation that has timed out. In addition the lines vary in thickness depending on the loading.

For ALL Nodes

• Click LAN Explorer
• Click Data Lists, Conversations.

A list of all the conversations appears.

By Protocol

Select Conversations using this Protocol.

• Click the Protocols button.
• Click the right mouse button on a Protocol.

Are Collisions Errors?

It is a commonly held misconception that collisions on an Ethernet LAN indicate problems. Collisions are a normal part of the operation of Ethernet LANs. In themselves, collisions are not errors. The busier the LAN the more likely the chance of a collision. The collision detection mechanism causes the nodes involved to cease transmission of data, send a jamming signal to warn other nodes not to transmit, and back off for a random time period before attempting to transmit again. All the time the carrier sensing (listening for signals on the cable) is continuing. A few collisions should be expected on a LAN operating normally.

Excessive collisions may be a symptom of trouble, or they may not. Very busy LANs are likely to have more collisions than lightly loaded LANs. If an Ethernet LAN becomes overloaded the throughput may drop dramatically due to nodes constantly having to back off and retransmit. This does not necessarily mean that there are faulty cards or devices on the LAN. Bridging or switching segments to limit the traffic may reduce the number of collisions.

Runt packets, Long packets (Jabbers), CRC and Frame Alignment Errors are of much greater significance when it comes to finding real problems.

What is Point & Shoot Tracing?

When you have identified a potential problem area 'Point and Shoot' operation provides a quick and easy way to isolate that Conversation from the others on the LAN and begin tracing it. Even on a small network the potential number of Conversations is huge. For example, a 10 node network can have up to 100 Conversations. The last thing you need is to have to stop, make a note of the addresses involved, exit the current program, run a separate tracing program and start to capture the relevant traffic.

How does it work?

Highlight the Conversation, or node on the LAN Map, or packet type, or an item on one of the LAN Explorer lists. Click the right mouse button and select Trace. Filters are automatically configured and the Packet Capture screen appears. As soon as a packet appears on the LAN matching the filter criteria it will be captured.

What Alarms does the product report?

Network Alarms are an important feature, but which Alarms are useful? What is a sensible value for each Alarm? Over how long a time period should Alarms run?

The currently supported alarms are:

• CRCs (Cyclical Redundancy Check) - A calculation performed on the data sent and received, verifying that there have been no corrupted packets or transmission errors.
• Short Packets (Runts) - Packets less than 64 bytes (60 bytes of data plus 4 bytes CRC) are short (runt) packets. They can be caused by faulty cables and plugs or by malfunctioning LAN cards.
• Long Packets (Jabbers) - Packets greater than 1518 bytes (1514 plus 4 bytes CRC) are faulty. Can be caused by faulty LAN cards.
• Broadcast Packets - Broadcast packets have a Destination Address of FFFFFFFFFFFF. They are often used by servers to advertise their services to all Nodes on the network, or to propagate routing information. Nodes, (on Novell NetWare), normally only send Broadcast packets when they log on.
• An abnormally high number of these packets may indicate a user continually logging out incorrectly (by switching the PC off) and logging in repeatedly. Since the Broadcast address is always a destination address it should never transmit, so the packets from counter should always be zero.
• Multicast Packets - Multicast packets are used to send data to groups of nodes. Nodes have to be configured to receive Multicasts.
• Alien Node - Alien Nodes are not part of your Address directory. You may be experiencing unauthorized users logging on to your LAN. If so their addresses would appear on your LAN and the addresses logged.
• LAN Over Usage - It is useful to know if your LAN is being swamped with traffic. The LAN Over Usage Alarm reports this. Of course only you will know what is an acceptable maximum, so the loading and time period are configurable.
• LAN Low Packet - count As with LAN Over Usage the minimum level and time period are user configurable.

You can also switch Alarms On and Off, set threshold values and time periods, keep a log of Alarms, and define an Action for each Alarm. Actions could include playing a WAV file or running another program such as a pager messaging system.

How do I identify the nodes on my network?

Identifying particular Nodes from their Ethernet Addresses used to be a big problem. The Traffic Monitor solves this problem elegantly and simply.

All devices on Ethernet are given unique addresses by their manufacturer. Manufacturers are assigned prefixes by the Institute of Electrical and Electronic Engineers. For example some 3Com devices begin with 00608C, other manufacturers have different prefixes, some manufacturers have more than one prefix. If you happen to know the prefixes than it would be possible to identify any particular make of device from its Ethernet Address. The software contains a comprehensive list of these prefixes built-in and will automatically replace the prefix with the actual manufacturers name, there is no need for you to remember them or to look them up. You can add new prefixes and manufacturer's names to the address directory.

So, for example an Address 00608C123456 would automatically be identified as a 3Com card, the Address appears as 3Com:123456. The rest of the address is left as a string of digits. This may well be enough for your needs, you may be able to identify Nodes simply from the make of their network cards. If you are using devices from manufacturers not included in the list, you can add their unique prefix to it.

However this is only part of the story.

Discovering Device Names Automatically

Click the Discovery button to begin detecting device names automatically. Discovery watches the network traffic, detecting device names as they appear on the wire. Discovery detects NetWare names, NetBIOS names and many others, IP addresses can often be resolved to names. Of course a single device may have multiple names, one for IP, one for NetBIOS, one for NetWare. You can select which name to use or Discovery can assign the names automatically.

Of course you can still edit any device name by hand if required so you can allocate devices to Groups or Types.

How do I add new addresses to my Address Directory?

The software automatically lists every address that appears on the network. The global address directory is called DTE.TXT and can contain up to 10000 addresses.

What about corrupt addresses?

If you have problems on the LAN which cause packets to be damaged you may see corrupt addresses appearing on your LAN map. Noise on the wire, faulty connections, out of spec cables, and so on can produce corrupt addresses. These are often easy to spot since they do not look like 'normal' addresses. Ethernet addresses usually begin with values such as 00, 02, 04, 06. Multicast addresses usually begin with 01, 03, 05, 07, 09. Corrupt addresses can begin with any value, AB, 12, AA, 55 etc. Corrupt addresses are usually characterized by only ever sending or receiving a single packet. You can see this clearly in the Statistics Output.

Rows of As or 5s in an address usually indicate 'streaming', caused when a faulty card sends a signal continuously on the wire. This is a potentially serious condition, as it will prevent other devices from sending any packets. Isolate the card by turning machines on or off until you can make the problem appear and disappear.

How do I get rid of corrupt addresses?

Clearly you do not need to keep any corrupt addresses, they simply take up space in your DTE.TXT file.

To remove corrupt addresses follow the steps below.

Step 1: Run the Traffic Monitor and open a segment map.

Devices not transmitting show up in black.

Step 2: Select Options, Remove Inactive Nodes.

You cannot do any harm by selecting this option. If you should happen to delete a device that really does belong on your LAN, but is currently turned off, it will simply reappear next time it logs on. Clearly you cannot remove nodes that are actually sending packets.

What causes Errors and what do they mean ?

There are many causes of errors on Ethernet and Token Ring LANs. They may occur when a packet has been transmitted by a faulty Network Interface Card, there may be a fault with the LAN cabling, a connector may be badly fitted, the cabling may be out of spec, there may be a protocol error. Packets can be malformed and therefore of no use.

Some error types generate more than one error report, and therefore the Total Errors counter may appear to be incorrect. For example, a Frame Alignment error appears as two errors because it will always generate a CRC error at the same time. Although the Error counter will show 1 CRC and 1 FA error, the Total Errors counter shows 1.

Many different types of Errors may be reported. On Ethernet:

• CRC Every packet contains a Cyclic Redundancy Check word built in. This is used to check that the packet has been transmitted correctly. Malformed packets fail the check and are reported as CRC errors. Some (but not all) CRC errors are also reported as FA errors.
• FA Frame Alignment error. An error where the bits that constitute a frame are incorrectly spaced (aligned). Frame Alignments always count as CRC errors as well.
• Short Packet Also known as Runt Packets. Packets less than 64 bytes long are always errors. In addition Short Packets may also be counted as CRC errors and possibly as FA errors as well.
• Long Packet Packets longer than 1518 bytes are always errors. In addition Long Packets may also be counted as CRC errors and possibly as FA errors as well.
• Software Missed On very busy networks it is possible that the software collection process may be overtaken by the network. The only guaranteed way to avoid this is by using (very expensive) dedicated hardware. Unlike most other manufacturers we report missed packets so that the statistics reporting will always return the correct values. A faster card and driver, or a more powerful computer will usually cope.
• Hardware Missed If the buffers in the card interface chip become overwhelmed an error is reported. Try a different make of card, some cope better than others on heavily loaded networks, or try using a later driver, normally available from the manufacturer's web site.
• Hardware Restarts If the card interface chip becomes confused an error is reported and the card resets itself. An unknown number of packets will be missed during the resetting process. Try a different make of card, some cope better than others on heavily loaded networks, or try using a later driver, normally available from the manufacturer's web site. Restarts may be also caused by cabling problems so check that too.

Note that Collisions are not normally considered as errors. Collision detection is part of the operation of an Ethernet network and you should expect them to happen from time to time on a perfectly healthy network. Excessive Collisions will result in the errors described above.

How do I print a list of addresses?

It will often be useful to print a list of MAC address details for the devices on your network.

Step 1: Run the Traffic Monitor

Step 2: Click the Output Statistics button.

Step 3: Click on Node List, Node Details

Step 4: Click on Notepad.

You have now made a text file of all the known details for all the nodes on your network.

Step 5: Save the file, renaming it as required.

Note that you can also print lists of nodes ranked in various ways, an error list, and a list of conversations.

How do I print out a list of IP to MAC address details?

You may need to get a list of which IP addresses against MAC addresses. Follow the steps below:

Step 1: Click on File, Create IP address file.

Step 2: Give the text file a name and optionally a new path.

Step 3: Save the file.

What is streaming?

A network card may fail in various ways. A common type of failure is when the card sends out a continuous signal on to the wire. This is known as streaming.

How does it appear in the Traffic Monitor?

Spurious Conversations appear in the Conversation Table, with unknown protocols. The addresses are very likely to appear as a string of A's or 5's, or a combination of the two. This arises because the Traffic Monitor sees packets with addresses (in binary) of 010101010101 or 101010101010 (555555 and AAAAAA in hexadecimal).

How do I find the offending card?

Switch nodes on (or off) or unplug them from the LAN one at a time until the problem addresses can be made to appear and disappear. Replace the suspect card and recheck.

What is spoofing?

ISDN bridges use 'spoofing' to fool each end of a link that the line is connected even when it isn't. Since many conversations are inactive for long periods, the only traffic is keep alives and/or acknowledgements which can be emulated by the bridge software. The benefits are obvious, if the line is not connected there is no charge. Although this system is usually quite reliable, problems can be very difficult to find; an analyzer that can trace conversations is indispensable.

How can I check if nodes are communicating with their intended Servers?

It is not very useful to have segment A's users talking to segment B's server and segment B's users talking to segment A' s server. Looking at the conversation table it is possible to assess whether a server is on the correct segment.

One Network was designed in an apparently logical way, but caused terrible congestion problems on a segment. The designer had 10 servers and several hundred users to connect up. The servers were designated for use by the accounts group, the marketing group and so on. The plan hatched: put all of the servers on one segment and bridge/route the users in. When the analyzer was put onto the segment, the average usage was 60%, with peaks of 90%. Ethernet does not work well over 50% usage and we recommend less than 30% as good rule of thumb. Furthermore, the bridges themselves became flooded and missed frames they were supposed to forward. The result was a network populated by users suffering poor response times.

How can I calculate network usage for cost accounting purposes?

Many organizations these days are run as groups of 'profit centers' and charge users on a time or packets basis for access to a server. When a WAN is involved, this can be very hard to administer.

The Traffic Monitor can collect statistics for Groups of devices and automatically calculate the bandwidth usage for each Group. To allocate nodes to Groups follow the steps below:

Step 1: Run the Traffic Monitor.

Step 2: Point the mouse cursor to the node that you want to add to a Group, click the right mouse button.

Step 3: Select Edit.

Step 4: Click on the down arrow beside the Group icon, or on the drop down list of names.

Step 5: Click on the required Group,

Step 6: Click OK.

You have now allocated the node to the named Group.

Adding a new Group or Type.

Step 1: Run the Traffic Monitor.

Step 2: Click on Options, Edit Types and Groups.

Step 3: Click on the + beside Groups, to add a new Group, or Types to add a new Type.

Step 4: Click on one of the Not Set entries.

Step 5: Type a new name for the Group or Type.

Step 6: Optionally select a bitmap to represent the new Group or Type.

Step 7: Click OK.

What is Tunnelling?

Tunnelling is a technique used to route packets even though the protocol in question is not designed to be routed. Typically it is used for protocols such as NetBIOS which

It is basically quite simple, for example, to use an IP Router to carry Novell Traffic the Server makes packets which have an IP layer slotted in before the IPX frame. The router can deal with these packets and deliver them accordingly, at the other end, the IP part is discarded and the packets is back to normal. Problems with this require the use of a Protocol Analyzer.

How do I get Protocol specific statistics?

Detailed Statistics are automatically collected on all the Protocols detected on the LAN. Access to the live data is available from several places in the software.

From the Traffic Monitor main screen

Step 1: Click on the Protocols Button.

A list of Protocols appears alongside a Pie Chart showing a breakdown by Packets Sent or by LAN Usage. Further details are also available.

Step 2: Select a Protocol from the list.

Step 3: Click the right mouse button, select Protocol Watch.

A statistics box appears showing live statistics for the selected Protocol. You can get a list of all the Conversations using the selected Protocol and you can also capture packets.

Detailed Statistics on NetWare, IP, TCP, UDP, ICMP

Step 1: Click on the Windows Menu Bar option, Statistics.

Step 2: Click on one of the common Protocols listed.

Select from NetWare, IP, TCP, UDP, or ICMP. A protocol specific statistics box appears giving a breakdown of packet types for the selected protocol.

From LAN Explorer

Step 1: Click the Explorer Button in the Traffic Monitor.

Step 2: Click the LAN Explorer Tab.

Step 3: Click the + beside Protocols for a list.

Step 4: Select a Protocol.

The statistics on the right of the screen are for the specified Protocol.

Step 5: Click the right mouse button on the Protocol for a menu of further options.

Alternatively click the Data Lists Tab, Protocols, for a list of Protocols with Packet and Octet counts.

Address Table limits in Routers and Hubs

Routers and Hubs hold internal tables of all reachable addresses, they use the list to decide the where to send incoming packets. In normal circumstances everything works fine.

However on a customer site recently it was noted that sections of the network were reported as being unreachable. The Traffic Monitor was seen to contain a series of ICMP error messages coming from a router with errors such as Network Unreachable and Host Unreachable. The actual networks and hosts were performing perfectly, the problem was with the router itself.

The Router, and this seems to be a common limitation, can hold a maximum of 1024 addresses. The network, although lightly loaded (3%), contained lots of multicast addresses coming in from remote segments, making a total of about 1400 addresses. The extra addresses were causing the Router problems and it was discarding the packets containing the unknown addresses.

The Traffic Monitor coped with the nodes easily and the Protocol Analyser picked up the error packets with no trouble.

A permanent solution to the problem is in the hands of the Router manufacturer.

How do I increase the number of Nodes that the software can handle?

When the software is issued it is set to handle up to 1000 Nodes. For most users this is more than enough. Of course there are networks with more than 1000 devices attached, but even on much smaller networks there are times when it may be necessary for the software to handle more than 1000 Nodes.

For example if a faulty card or cable is damaging packets the addresses might appear as unknown Nodes on the network, and quickly exceed the 1000 Node limit. See Application Note 8 for details of addressing. To increase the limit, follow the steps below.

Step 1: Using Notepad, open the file called PODLINK.INI.

PODLINK.INI contains all the variables that the user can reset. PODLINK.INI is automatically created when the software is first loaded. PODLINK.INI will be found in the subdirectory where you installed your product.

Note: In versions earlier than version 5.0 the file was called WINPOD.INI. Overwriting an older installation will cause the WINPOD.INI file to be removed and the contents copied to PODLINK.INI.

Step 2: You may change any of the values in PODLINK.INI.

Step 3: Edit the entry MaxConvs in the section marked [LocalPod].

The number here specifies the Maximum Number of Conversations that the software can handle. The value is always 3 times the number of Nodes.

For example, MaxConvs=3000 allows 1000 Nodes, MaxConvs=9000 allows for 3000 Nodes. The maximum permitted value for MaxConvs is 30000, equivalent to 10000 Nodes.

Step 4: Save the file and Exit Notepad.

Checking your new settings

To check your new settings follow the steps below.

Step 1: Run the Traffic Monitor software.

Step 2: Open a segment and view the LAN map.

Step 3: Click on Statistics on the Windows Menu Bar at the top of the map.

Step 4: Click on System.

The Maximum Conversations and Node boxes will show the current settings.

Updating to Windows98

When you are installing a new version of the software over an existing one in Windows 98 follow the steps below. The installation is completely normal, only the driver loading at the end of the installation is affected.

Step 1: Inset the CD as normal.

Step 2: Follow the loading instructions.

Step 3: When the Network Control panel and the Help appears close them both.

There is no need to reload the card driver, it will have been updated automatically as part of the installation process.

Repairing Links in the Excel Macro

Report Generator makes a document containing the links to the CHARTS.XLS file. If these links become corrupted or changed the Excel diagrams will not appear.

Step 1: Click on the error message for a faulty link.

Step 2: Click on Edit, Links…

Step 3: The chart relating to the broken link will be highlighted automatically.

Step 4: Click on Change Source…

Step 5: Search for the file CHARTS.XLS.

By default this will be on C:\PROGRAM FILES\CHEVIN\CNApro.

Step 6: Click on CHARTS.XLS.

Step 7: Click on Open.

Step 8: Click on Update Now.

Step 9: Click OK.

Step 10: Click on File, Save to keep the changes you have made.

Saving a Report Document

Report Generator makes a document containing the links to the CHARTS.XLS file. To save the report follow the steps below.

Step 1: In the Word document click on Edit.

Step 2: Click on Links…

Step 3: Click on the first item in the list, hold Shift and drag the scroll bar to the bottom of the list.

All the items on the list will be highlighted.

Step 4: Click on Break Link.

Step 5: Save the document using a new name.

There is no need to save the ETHERNET REPORT.DOC. It will be overwritten the next time you generate a report.

There is no need to save any of the Excel diagrams. Just answer No to the messages about saving the files.

Counting known TCP and UDP Ports

Specifying a Port Number

Adding an entry to the Statistics Box is very easy. You may specify a value for the port and which Statistics Box you want it to appear in. In addition if you want the value to be replaced by a name or piece of text that too may be specified. The steps are outlined below.

Step 1: Make a file to hold the required values.

The file names are fixed. To add entries to the TCP Statistics Box use a text editor (Notepad) to make a file called TCPPCTR.TXT.

To add entries to the UDP Statistics Box use a text editor (Notepad) to make a file called UDPPCTR.TXT.

Step 2: Type in the Port Numbers that you want to monitor.

Just add the Port Numbers you want to monitor. End each line by pressing the Enter key.

Step 3: Save the file.

If you want the new Port Number to appear in the Statistics box as a piece of text follow the steps below.

Replacing the Port Number with Text

Step 1: Open the file called UDPPORT.TXT or TCPPORT.TXT using a text editor (Notepad).

These files contain lists of defined UDP or TCP ports.

Step 2: Type in the text that you want to see followed by a comma and the new Port Number.

Step 3: Save the file.

Step 4: Run the Traffic Monitor, open the Statistics boxes for UDP or TCP.

The new Port Number will be added to the Statistics box.

What does a blue line or blue Node mean?

All Nodes and Conversations are color coded to show their current status.

Blue lines indicate that a Conversation is single sided, for example to a Broadcast or Multicast address. Broadcast or Multicast addresses will not transmit packets.

A blue node indicates that the device has only ever been seen as a destination address on the network, it has never transmitted a packet

What does a dotted line mean?

All Nodes and Conversations are color coded to show their current status.

A dotted line appears when a Conversation times out.

Users may set the Node and Conversation timeout.

Step 1: Click on Options on the Windows Menu bar.

Step 2: Click on Set Global Options.

Step 3: Type a number of minutes in the Node Idle or Conv Idle boxes.

Note that on the LAN Map the Node will appear in outline indicating that the device has timed out.

Why do my Token Ring addresses and my MAC Table appear with all zeros?

There are two different Token Ring adapter chipsets in use. Only those using the Texas Instruments chips are capable of being used in 'Promiscuous' receive mode, that is they will receive all the frames on the ring.

If your addresses are all zeros and the MAC Table is empty or grayed out you are attempting to use a card that is not in promiscuous receive mode. Check in the card Properties for an option to switch Statistics Gathering (or Monitor Mode) on.

We recommend Madge 16/4 Smart Ring Node cards. There is an option to switch on Statistics Gathering Mode and this MUST be selected.

NB: Madge Presto cards are NOT suitable for use with the analyzer.

IBM and 3Com cards are NOT suitable for use with the analyzer.

Why does my Traffic Generator percentage jump to 100%?

If the Traffic Generator program is left running it may occasionally jump to 100% usage. This is due to the use of a Screen Saver on the PC. While the Screen Saver is active the Traffic Generator is unable to produce traffic. When it does get access to the processor the Traffic Generator attempts to bring the usage back to the preset level by producing as many packets as possible in the shortest time. The packets generated rises.

To prevent this disable the Screen Saver.

Why is my screen slow to redraw on Larger Networks?

On larger networks, that is greater than 500 nodes, if the PC is slow or has a limited amount of RAM, you may find it impossible to update the LAN Map and service mouse clicks at the same time. The cursor will alternate between the egg timer and the cursor, often to the point where the mouse click will not be responded to.

The easiest thing to try first is adjusting the time between Pod polls. This will sample the statistics less often and give the PC time to redraw the LAN Map. Follow the steps below.

Increasing the Time between Pod Polls

Step 1: Run the Traffic Monitor and open the LAN Map for your segment.

Step 2: Click on Options, Set Global Options.

Step 3: Increase the value in No. of Seconds between Pod Polls.

Try increasing the value to 10. If the problem reoccurs increase the value further. Default value is 2 seconds, maximum is 60 seconds.

Switch Off the Action Bars

In addition switching off the Action Bars that appear underneath the node icons will speed up the screen redraw.

Step 1 Run the Traffic Monitor and open the LAN Map for your segment.

Step 2 Click on Options, Set Global Options.

Step 3 Untick the box next to Display Action Bars.

General Recommendations

• On large networks, particularly is usage is heavy, always use the fastest possible PC, 233Mhz or faster Pentium, with the maximum amount of RAM, 64M or more.
• Always use the latest card driver supplied by the manufacturer. Most manufacturers keep downloadable drivers on their websites.
• Close any other Applications unless they are essential.

If remote Podlinks are connected over slow links increase the Time between Pod Polls option as described above.

Why do I get a Macro Error 1004?

If the Excel macro fails it often reports an Error 1004 message, such as:

Error 1004 – [ Method Name ] method of [ Object ] class failed

This message is a general error report and can mean that there has been a failure reading or writing from a file.

A hardware failure may have arisen or corrupted files may be responsible for this error. Check that the CSV files are present in the correct directory, normally in the Program Files/Product Name/Segment Name directory.

Possible additional reasons include the following:

• An argument contains a value that isn’t valid. A common cause of this problem is an attempt to access an object that doesn’t exist. For correct operation of macros a minimum of three workbooks must be open. The minimum workbook setting of three is the standard default setting from within Microsoft Excel.

If this error occurs, then look at the following:

Step 1: In Excel, go to Tools

Step 2: Go to Options

Step 3: Go to General

Step 4: Look at Sheets in New Workbook

Step 5: Ensure that a value of 3 or more is set

Step 6: Check all CSV files and the SEGMENTS.TXT file to ensure that all cells are filled correctly. I.e. All the data is separated into individual cells and not bunched into one single cell

Step 7: Run the Report Generator again and try the macro again.