Connections Tab

Usage: Click the Help button at any time to get online help for the settings in the current tab.

This opens a World Wide Browser, such as Netscape Navigator or Internet Explorer, and displays information about the settings.

To close the help display, close the browser as you would any other program. For example, in Windows 95, click the X button at the upper right of the window or choose Exit from the File menu.

With some browsers, a new window will open each time you click the Help button. If this happens, simply close each window when you're done with it.

Finding setting descriptions: At the beginning of each help window is a table of contents for the window. To go information about a particular setting, click the hyperlink (the underlined text in most browsers) for that setting in the contents. To return to the contents, use the control in your browser for going back, such as the Back button in Netscape Navigator or Internet Explorer 3.0.

Usage: Click this button to send the settings shown in the Java-Based Pipeline Configurator to the Pipeline.

These new settings replace the previous settings in the Pipeline. You can use this button at any time; you do not have to enter values for all settings before updating the Pipeline.

Usage: Click this to save the settings shown in the Java-Based Pipeline Configurator in a file on your computer.

You then use your operating system's facility for specifying a name and location for the file, such as the Save To dialog box in Windows 95.

You can later open a file in which you've saved settings by clicking the Open a Configuration File button in the first window that appears when you run the Java-Based Pipeline Configurator.

You can use this button at any time; you do not have to enter values for all settings before saving settings.

Usage: Click this to open the Pipeline Monitor.

When you click the Connections tab, a list of current connection profiles, if any, appears. To select a profile, click it. There are also three buttons on the display:

• New

  Click this to create a new connection profile. This displays the Edit Settings window, which contains the settings for the profile.

• Edit

  Click this to change settings for the selected profile. This opens the Edit Connections window. This displays the Edit Settings window, which contains the settings for the profile.

• Delete

  Click this to delete the selected profile.

Usage: Enter the name or MAC address of the remote device.

You can enter up to 31 characters.

The value you specify is case sensitive, and must exactly match the name of the remote device. If you are not sure about the exact name, contact the administrator of the remote network.

Dependencies: Keep this additional information in mind:

• The Station setting appears in the list of Connection Profiles display when you click on the Connections tab; however, if you leave the setting blank, the IP Address setting appears instead in the Connection Profiles display.

• The remote device that the Station setting specifies is the device actually placing or answering the call; it is not necessarily the same as the source or destination of packets using the link.

• The Pipeline does not currently use the Domain Name System (DNS) to determine the IP address of the device specified by the Station setting.

• When the Pipeline receives a PPP or MP+ call from an Ascend unit, it tries to match the caller's Name to the value of the Station setting in some Connection Profile.

  If the Pipeline finds a match and authentication is turned on, the Pipeline then tries to match the caller's Password value to the Password value in that same Connection Profile.

Usage: When the box is checked, the profile is available for use.

Dependencies: Keep this additional information in mind:

• If the profile is active, the static route implied by the Connection Profile can appear in the static routing table.

• If you deactivate a Connection Profile, the static route implied by the Connection Profile is removed from the static routing table.

Description: This checkbox enables or disables protocol-independent bridging for a call. If you disable bridging, you must enable routing by checking Enable IP Routing or Enable IPX Routing in the Protocols tab.

Usage: Check the box to enable bridging.

This is unchecked by default.

Dependencies: The effect of the Enable Bridging setting depends upon how you set the IP settings (providing IP settings enables IP routing) and Enable IPX setting in the Protocols tab.

Enable Bridging and Enable IP Routing

• If the Enable Bridging setting is checked and IP routing is enabled, the Pipeline routes IP packets and bridges all other packets.

• If the Enable Bridging setting is checked and IP routing is disabled, the Pipeline bridges all packets.

• If the Enable Bridging setting is not checked and IP routing is enabled, the Pipeline routes only IP packets.

• If the Enable Bridging setting is not checked and the IP routing is disabled, an error occurs and you cannot save the profile.

You must enable bridging or routing, or both.

Enable Bridging and Enable IPX Routing

• If both the Enable Bridging and Enable IPX Routing setting are checked, the Pipeline routes IPX packets and bridges all other packets.

• If the Enable Bridging setting is checked and the Enabled IPX setting is not checked, the Pipeline bridges all packets.

• If the Enable Bridging setting is not checked and the Enabled IPX setting is checked, the Pipeline routes only IPX packets.

• If the Enable Bridging setting is not checked and the Enabled IPX setting is not checked, an error occurs and you cannot save the profile.

You must enable bridging or routing, or both.

Additional Dependencies

• Bridging must be enabled on both the dialing and answering sides of the link.

  The Enable Bridging setting must be checked in both the Connection Profile on the dialing side and the Answer tab on the answering side. Otherwise, the Pipeline does not bridge the packets.

• Enable Bridging in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Enable Bridging setting in the Connection Profile takes precedence.

Usage: Choose one of the following:

• Switched
  This setting specifies a link that consists entirely of switched channels.

  The Switched setting is the default.

• Nailed
  This setting specifies a link that consists entirely of nailed-up channels.

  You must use the Group setting to specify which channels are in the connection.

• Nailed/MPP
  This setting specifies a link that consists of both nailed-up and switched channels. The Pipeline establishes this connection whenever any of its nailed-up or switched channels are connected end-to-end. If a Nailed/MPP link is down and the nailed-up channels are down, the link cannot re-establish itself until the Pipeline brings up one or more of the nailed-up channels, or dials one or more switched channels.

  Typically, the switched channels are dialed when the Pipeline receives a packet whose destination is the unit at the remote end of the Nailed/MPP connection. The packet initiating the switched call must come from the caller side of the connection.

  If a channel in a call fails for any reason, and the total number of channels in the Nailed/MPP connection falls below the value of the Min Channels setting (displayed when you click the Bandwidth button), the Pipeline tries to add a switched channel to bring the connection back up to the minimum.

  If a failed channel is in the group specified by the Group setting, that channel is replaced with a switched channel, even if the call is online with more than the minimum number of channels. Failed nailed-up channels are replaced by switched channels, regardless of the Min Channels setting.

• Permanent Switched
  This setting specifies a permanent switched connection.

  A permanent switched connection is an outbound call that attempts to remain up at all times. If the unit or central switch resets, or if the link is terminated, the permanent switched connection attempts to restore the link at ten-second intervals.

  Use this setting if your telephone company charges for each incoming and outgoing connection attempt, but does not charge for connection time on local calls. Ascend's regular bandwidth-on-demand feature conserves connection time but causes many connection attempts. A permanent switched connection performs the opposite function--it conserves connection attempts but causes a long connection time.

  For the answering device at the remote end of the permanent switched connection, you should configure the Connection Profile to answer calls but not originate them. If the remote device initiates a call, the Pipeline simply does not answer it. This situation could result in repeated charges for calls that have no purpose. To keep the remote device from originating calls, set Dial In/Out to Dial In Only for that device.

• The Pipeline determines the minimum number of channels by the value of the Min Channels setting or the number of nailed-up channels in the group, whichever is greater.

  The Pipeline does not count a nailed-up channel that is unused.

• The Pipeline adds or subtracts switched channels on a Nailed/MPP connection as required by the settings on either side of the connection.

  Each side makes its calculations based on the traffic received at that side. If the two sides of the connection disagree on the number of channels needed, the side requesting the greater number prevails.

• The DO Hangup setting works only from the caller side of the connection when you choose Nailed/MPP.

• The Idle Setting works for both sides of the connection when you choose Nailed/MPP. However, if the answering side of the connection brings the link down because of an Idle timeout, the calling side can bring it back up.

• If the link consists entirely of switched channels (Call Type is Switched), the Group setting does not apply (Group is dimmed).

• In a Connection Profile, the encapsulation must be MPP (Encaps is MPP) in order to set Call Type to Nailed/MPP.

• When you set Call Type to Permanent Switched in a Connection Profile, Dial In/Out is dimmed and does not apply because permanent switched connections are always outbound.

Usage: Enter a number between 0 and 65535. If you specify 0 (zero), Pipeline does not enforce a limit; an idle connection stays open indefinitely.

The default setting is 120 seconds.

Dependencies: Keep this additional information in mind:

• In the Answer tab or Connection Profile, Idle Timeout does not apply to nailed-up links; that is, the Idle Timeout setting is dimmed when the Call Type setting is Nailed.

• If MP+ encapsulation is in use and the bandwidth utilization on both sides of the connection drops below the value entered for the Idle % setting (in the Bandwidth setting of Connections tab), the Pipeline clears the call, regardless of the value you enter for the Idle Timeout setting.

• Idle Timeout in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Idle Timeout setting takes precedence.

• If the Require Connection Profile setting is checked in the Answer tab, Idle Timeout does not apply (Idle Timeout setting is dimmed) in the Answer tab.

• Because the Idle % setting is dependent on traffic levels on both sides of the connection, we recommend that you use the Idle Timeout setting in preference to it.

To make a 64 Kbps data connection, all the telephone companies used for the connection must provide 64 Kbps service. If any of them provide only 56 Kbps service, you must specify a 56 Kbps connection.

In a few cases, it is necessary to use a voice connection to carry digital data. This technique, known as Data over Voice (DOV), also can be used when voice calls are less expensive than data calls.

Usage: Choose one of the following:

• 56K
  This setting specifies a 56 Kbps data connection.

• 56KR
  This setting specifies a restricted 56 Kbps data connection.

• 64K
  This setting specifies a 64 Kbps data connection.

• Voice
  This setting specifies a Data over Voice (DOV) connection, which uses a voice call to carry digital data. The Voice setting applies only to switched channels.

Note: Although 56KR service carries data at the same rate as 56K service, 56KR and 56K are different types of service. If the network to which you're connecting requires 56KR service, Link Type must be set to 56KR.

Dependencies: Keep this additional information in mind:

• Data over Voice (DOV)--the data transmission technique used when the value of Link Type is Voice--is often less reliable than a standard data connection. To avoid loss or corruption of data when using Data over Voice, do the following:
  • Make sure that the connection carries only digital data.

  • Make sure that the phone company is not using any intervening loss plans to economize on voice calls.

  • Do not use echo cancellation; techniques for removing echoes from analog calls can corrupt data

• The speed of DOV connections is limited to 56 Kbps.

• If the Pipeline requests a type of service that is unavailable, it cannot connect to the network service provider.

• The speed of a data connection is the speed, in kilobits per second, specified by this setting (56 Kbps if the value is Voice) multiplied by the number of B channels (either 1 or 2) used for the connection.

Usage: Choose one of the following:

• None
  This turns off data compression.

• Stac
  This turns on data compression.

  The Pipeline applies the STACKER LZS compression/decompression algorithm. STAC is the default.

• Microsoft Stac
  This turns on Microsoft LZS Coherency Compression for Windows95, a proprietary compression scheme for Windows 95 only (not for Windows NT).

  Both sides of the link must set Compression to Stac or Microsoft Stac to turn on data compression.

• Stacker LZS Compression (as defined in the Internet Draft of November 1995) and Microsoft LZS Coherency Compression for Windows 95 both use the same PPP option to indicate that their compression scheme is in use. Therefore, a router can have difficulty determining exactly which compression method a caller is requesting. Ascend units handle this ambiguity in the call by always using the compression scheme specified in the Connection Profile; if there is no Connection Profile, the Ascend unit uses the compression scheme specified in the Answer tab.

  If the caller requests Microsoft Stac and the profile does not specify Microsoft Stac compression, the connection seems to come up correctly, but no data is routed. If the profile is configured with Microsoft Stac and the caller does not acknowledge that compression scheme, the Pipeline attempts to use standard Stac compression. If it cannot use standard Stac compression, it uses no compression at all.

• Both sides of the link must set Compression to Stac or Microsoft Stac to turn on data compression.

• The Compression setting applies only if the link uses PPP encapsulation (Encapsulation setting is PPP or MPP).

  When the Encapsulation setting is MPP, both the dialing side and the answering side of the link must support MP+. If only one side supports MP+, the connection uses MP or standard single-channel PPP. When the Encapsulation setting is PPP, the connection uses only PPP.

• Compression in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Compression setting takes precedence.

Usage: Click the Data Filter button. In the window that appears, select a data filter and then click OK.

If you choose None, the Pipeline forwards all data packets. None is the default.

Dependencies: Keep this additional information in mind:

• The Pipeline applies a Call Placement Filter after applying a data filter; only those packets that the data filter forwards can reach the Call Placement Filter.

• If IPX client bridging is in use (Discard RIP, SAP setting is Client), set Data Filter to None.

• Do not confuse this Data Filter setting with the Data Filter setting in the Connections tab.

  This Data Filter setting filters data packets on the Pipeline unit's local LAN interface; the Data Filter setting in the Connections tab filters data packets on the Pipeline unit's WAN interface. The WAN interface is the port on the Pipeline that is connected to a WAN line.

• This Data Filter setting applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Data Filter setting takes precedence.

• If the Require Connection Profile setting is checked in the Answer tab, Data Filter does not apply (Data Filter setting is dimmed) in the Answer tab.

Usage: Enter a number between 1 and 3.

Example: If Call Type is Nailed/MPP in a Connection Profile, the setting 3 assigns one nailed-up group to the profile.

Dependencies: Keep this additional information in mind:

• The Group setting does not apply (Group is dimmed) if the link consists entirely of switched channels (Call Type is Switched).

• If you add channels to the Group setting and save your changes, the Pipeline adds the additional setting to any online connection that uses the group.

• Do not assign more than one active Connection Profile to a group.

• Do not assign a Connection Profile to a group that a Frame Relay Profile uses.

• If you are using an ISDN BRI line the Pipeline assigns the B channels to the following groups:
  • 1 represents the B1 channel

  • 2 represents the B2 channel

Usage: Choose one of the following:

• Dial In Only
  This specifies that the Pipeline can receive calls from the destination specified in the Connection Profile, but cannot initiate calls to that destination.

• Dial Out Only
  This specifies that the Pipeline can dial out to the destination specified in the Connection Profile, but cannot answer calls from that destination.

• Both
  This specifies that the Pipeline can initiate calls to the destination specified in the Connection Profile, and that it can receive calls from that destination as well.

  Both is the default.

Usage: Choose one of the following:

• MPP
  MP+ (Multilink Protocol Plus) extends the capabilities of MP (Multilink PPP) to support inverse multiplexing, session management, and bandwidth management. MP is an extension of PPP that supports the ordering of data packets across multiple channels.

  MP+ consists of two components: a low-level channel identification, error monitoring, and error recovery mechanism, and a session management level for supporting bandwidth modifications and diagnostics. MP+ enables the Pipeline to perform Dynamic Bandwidth Allocation (DBA)-that is, MP+ enables the Pipeline to add or remove channels without disconnecting a link as the need for bandwidth increases or decreases.

  Both the dialing side and the answering side of the link must support MP+. If only one side supports MP+, the connection then tries to use MP. If that fails, the connection uses standard single-channel PPP. Note that neither MP nor PPP support DBA.

• PPP
  PPP (Point-to-Point Protocol) provides a standard means of encapsulating data packets over a single-channel WAN link that a Connection Profile sets up. It ensures basic compatibility with non-Ascend devices.

  For this setting to work, both the dialing side and the answering side of the link must support PPP.

• MP
  Multilink PPP (MP) is an extension of PPP that supports the ordering of data packets across multiple channels and inverse multiplexing.

• Frame Relay
  Frame Relay is an HDLC-based packet protocol that enables you to send data to a destination using one or more frame relay switches within a private network or a public carrier's network. HDLC stands for High Level Data Link Control.

  From the viewpoint of the Pipeline, a frame relay switch is an endpoint for all DLCIs (Data Link Connection Indicators) connecting to it. A DLCI identifies a Connection Profile as a logical link. The frame relay switch connects the endpoints of the DLCIs to each other to make a virtual permanent circuit to which users can connect. The circuit acts like a wire between two endpoints with a fixed maximum bandwidth.

• The Encapsulation setting does not apply (Encapsulation setting is dimmed) when the Pipeline answers a call, or if the link consists of only nailed-up channels (Call Type setting is Nailed).

• If the Call Type setting is Nailed/MPP then Encapsulation must be set to MPP. In this case, or whenever the Encapsulation setting is MPP, the Pipeline adds or subtracts switched channels on the connection as required by the DBA settings on either side of the connection.

  DBA, or Dynamic Bandwidth Allocation, enables the Pipeline to use average line utilization (ALU) of transmitted data as the basis for adding or subtracting bandwidth from a switched connection without terminating the link. MP+ and AIM support Dynamic Bandwidth Allocation. Each side makes its calculations based on the traffic received at that side. If the two sides of the connection disagree on the number of channels needed, the side requesting the greater number prevails.

• If the Encapsulation setting is MPP and the Call Type setting is Nailed/MPP, the minimum number of channels in the link is the number set by Min Channels or the number of nailed-up channels in the group, whichever is greater.

• If the Encapsulation setting is MPP and the Call Type setting is Nailed/MPP, the maximum number of channels in the link is the number set by Max Channels or the number of nailed-up channels in the group, whichever is greater.

  The Pipeline does not count a nailed-up channel that is not online.

Usage: The default setting is 1524; you should accept this default unless the device at the remote end of the link cannot support it.

If the administrator of the remote network specifies that you must change this value, enter a number between 1 and 1524.

Dependencies: Keep this additional information in mind:

• The Max PPP Packet setting applies to any link using PPP encapsulation (Encapsulation setting is MPP, MP, or PPP).

  When the Encapsulation setting is MPP, both the dialing side and the answering side of the link must support MP+. If only one side supports MP+, the connection uses MP or standard single-channel PPP. When the Encapsulation setting is PPP, the connection uses only PPP.

• Max PPP Packet Size in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Max PPP Packet Size setting takes precedence.

• If Require Connection Profile setting is checked in the Answer tab, Max PPP Packet Size does not apply (Max PPP Packet Size setting is dimmed) in the Answer tab.

Usage: Enter the password.

You can enter up to 20 characters; the password is case sensitive. Leave the field blank if the remote end does not require a password.

Dependencies: Keep this additional information in mind:

• The far end must specify a value for Password when the link uses PPP encapsulation (Encapsulation setting is PPP or Encapsulation setting is MPP) and the Pipeline uses PAP, CHAP, or Cache-Token authentication (Authentication setting is PAP, Authentication setting is CHAP, or Authentication setting is Cache-Token).

  When the Encapsulation setting is MPP, both the dialing side and the answering side of the link must support MP+. If only one side supports MP+, the connection uses MP or standard single-channel PPP. When the Encapsulation setting is PPP, the connection uses only PPP.

When you enable the callback feature, the Pipeline hangs up after receiving an incoming call that matches the one specified in the Connection Profile. The Pipeline then calls back the device at the remote end of the link using the Number setting specified in the Connection Profile.

You can use this setting to tighten security, as it ensures that the Pipeline always makes a connection with a known destination.

Usage: Check this to enable the callback feature.

By default, this is not checked.

Dependencies: Keep this additional information in mind:

• The Force Callback setting does not apply (Force Callback setting is dimmed) if all channels of the link are nailed up (Call Type setting is Nailed).

• If Force Callback is checked, you must also set Dial In/Out to Both, because the Connection Profile must both answer the call and call back the device requesting access.

  By the same token, any device calling into a Connection Profile set for callback must be configured to both dial calls and answer them.

LQM is a feature that enables the Pipeline to monitor the quality of a link. LQM counts the number of packets sent across the link and periodically asks the remote end how many packets it has received. Discrepancies are evidence of packet loss and indicate link quality problems.

LQM causes the generation of periodic link quality reports. Both ends of the link exchange these reports.

Usage:

• If the checkbox is checked, the Pipeline requests LQM.

• If the checkbox is not checked, the Pipeline does not request LQM.

  This is unchecked by default.

• Both sides of the link negotiate the interval between periodic link quality reports; however, the interval must fall between the minimum interval (as set by Minimum LQM) and the maximum interval (as set by Maximum LQM).

• If LQM is turned off [Request Link Monitoring (LQM) setting is unchecked], the Maximum LQM and Minimum LQM settings do not apply (Maximum LQM setting is dimmed and Minimum LQM setting is dimmed).

• LQM applies only if the Encapsulation setting is PPP.

• Request Link Monitoring (LQM) in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Request Link Monitoring (LQM) setting takes precedence.

Usage: Enter a number between 0 and 600. The default is 600.

Dependencies: Keep this additional information in mind:

• If the Request Link Monitoring (LQM) setting is unchecked, the Minimum LQM setting does not apply (Minimum LQM setting is dimmed).

• Minimum LQM in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the Minimum LQM setting in the Connection Profile takes precedence.

Usage: Enter a number between 0 and 600. The default is 600.

Dependencies: Keep this additional information in mind:

• If the Request Link Monitoring (LQM) setting is unchecked, the Maximum LQM setting does not apply (Maximum LQM setting is dimmed).

• Maximum LQM in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the Maximum LQM setting in the Connection Profile takes precedence.

Usage: Enter the calling party's phone number. You can enter up to 20 characters.

Usage: Enter the number to be used for Called Number authentication.

Example: 5551234

Usage: Enter a telephone number. You can enter up to 37 characters, and you must limit those characters to the following:

1234567890()[]!z-*#|

The Pipeline sends only the numerical characters to place a call.

Dependencies: Keep this additional information in mind:

Number does not apply (Number setting is dimmed) when all channels are nailed up (Call Type setting is Nailed) or if you are using frame relay encapsulation (Encapsulation setting is Frame Relay).

Usage:

• If the checkbox is checked, the Pipeline dials a link if (a) the link is not online and (b) the Pipeline receives a frame whose MAC address is set to broadcast.

  When a device on the local Ethernet interface sends out broadcast packets that the Pipeline must bridge to another network, the Pipeline starts up a session for each Connection Profile in which the Dial on Broadcast setting is checked. Gradually, it builds an internal bridge table based on experience; this table helps to limit the number of calls by recording the appropriate destination network for various addresses.

• If the checkbox is unchecked, broadcast packets do not initiate dialing.

  If you choose this setting, the Pipeline relies on its bridge tables (set with the Bridge Table setting), which contain remote physical addresses you have manually entered.

Your carrier determines the billing number, and uses it to sort your bill. If you have several departments, and each department has its own Bill-to Number, your carrier can separate and tally each department's usage.

Usage: Enter the telephone number to be billed. You can specify up to ten characters, and you must limit those characters to the following:

1234567890()[]!z-*# |

Dependencies: Bill-to Number for outgoing calls on an ISDN BRI line applies only to installations in Australia.

Usage: Choose one of the following:

• None
  This specifies that the Pipeline does not request an authentication protocol for outgoing calls.

  None is the default.

• PAP (Password Authentication Protocol)
  This is a PPP authentication protocol.

  PAP provides a simple method for the Pipeline to establish its identity in a two-way handshake. Authentication takes place only upon initial link establishment, and does not use encryption.

  If you choose PAP, the Pipeline requests this protocol for authentication. The remote device must support PAP.

  Note that if you choose this setting, the Pipeline requests PAP authentication but will use CHAP authentication if the called unit requires CHAP. Choose this setting for non-token card authentication if you would allow sending your password unencrypted.

• CHAP (Challenge Handshake Authentication Protocol)
  This is a PPP authentication protocol. If you choose CHAP, the Pipeline requests this protocol for authentication. The remote device must support CHAP.

  CHAP is more secure than PAP. CHAP provides a way for the remote device to periodically verify the identity of the Pipeline using a three-way handshake and encryption. Authentication takes place upon initial link establishment; a device can repeat the authentication process any time after the connection is made.

  Note that if you choose this setting, the Pipeline will not bring up the connection using PAP. Choose this setting for non-token card authentication if you do not wish to send your password unencrypted; that is, if you do not wish to be authenticated through PAP.

• PAP-Token
  This is an extension of PAP authentication. It requires the following:
  • The Network Access Server (NAS) must be running the Ascend RADIUS daemon.

  • There must be a RADIUS profile that matches the caller's name.

  • The RADIUS profile must access an ACE or SAFEWORD authentication server.

  In PAP-Token, the user making outgoing calls from the Pipeline authenticates his or her identity by entering a password derived from a hardware device, such as a hand-held security card. The Pipeline prompts the user for this password, possibly along with a challenge key. The NAS obtains the challenge key from a security server that it accesses through RADIUS.

  RADIUS (Remote Authentication Dial In User Service) is a protocol by which users can have access to secure networks through a centrally managed server. You can store virtually all Connection Profile information on the RADIUS server in a flat ASCII database.

• PAP-Token-CHAP
  This is nearly identical to PAP-Token. It requires the following:
  • the NAS must be running the Ascend RADIUS daemon

  • there must be a RADIUS profile that matches the caller's name

  • the RADIUS profile must access an ACE or SAFEWORD authentication server

  • the user profile must specify an auxiliary password (Ascend-Receive-Secret) that matches the MPP Password setting in the Connection Profile.

  Note: If MPP Password and Ascend-Receive-Secret do not match, it does not prevent the initial connection from succeeding, but the Pipeline cannot extend an MP+ call beyond a single channel.

  In all authentication protocols, including PAP-Token and PAP-Token-CHAP, the Pipeline individually authenticates all channels of an MP+ call. If the answering unit requires security card authentication, PAP-Token and PAP-Token-CHAP begin identically when authenticating the first channel of an MP+ call. However, when the Pipeline adds additional channels to the MP+ call, PAP-Token requires security-card authentication for each new channel, while PAP-Token-CHAP uses CHAP authentication for all new channels. CHAP authentication works automatically, without the use of a hand-held security card.

• Cache-Token
  This begins authentication using a hand-held security card, and fills a token cache set up for you on the RADIUS server at the remote site. It requires the following:
  • The NAS must be running the Ascend RADIUS daemon.

  • There must be a RADIUS user profile that matches the caller's name.

  • The RADIUS user profile must access an ACE or SAFEWORD authentication server.

  • the RADIUS user profile must specify an auxiliary password (Ascend-Receive-Secret) that matches the Password setting in the Connection Profile and defines Ascend-Token-Expiry in its first line.

  Note that if Password and Ascend-Receive-Secret do not match, it does not prevent the initial connection from succeeding, but subsequent connections (specifically, disconnecting/reconnecting or adding channels) fail until the cached token expires.

  CHAP authenticates your subsequent calls without using your hand-held security card. After a period of time configured in your entry in the RADIUS users file, the token cache expires and the next call you place must again be authenticated using your hand-held security card.

• The link must use PPP or MP+ encapsulation (Encapsulation setting is PPP or Encapsulation setting is MPP).

• If you request PAP or CHAP, you must also specify a password using Password in a Connection Profile.

• On a nailed-up link (Call Type setting is Nailed), you must set Authentication to the same value at both ends of the connection; that is, Authentication at the local and remote ends must all contain the same value.

• PAP-Token and PAP-Token-CHAP require configuration of a SAFEWORD or ACE entry in the NAS's RADIUS users file with the caller's name.

• For information on prompting for a password at a host, see the APP Server, APP Host, and APP Port settings.

• Dial on Broadcast must be enabled when a PC on the same Ethernet as the Pipeline runs APPSRVR1 or APPSRVR2 to open a connection protected by security-card authentication.

Usage: Enter a password. You can enter up to 20 characters; the password is case sensitive. The default is null.

Dependencies: Keep this additional information in mind:

• If the Authentication setting is None, the Password setting does not apply (Password setting is dimmed).

• You must specify a value for Password when the link uses PPP encapsulation (Encapsulation setting is PPP or MPP) and the Pipeline uses either PAP or CHAP authentication (Authentication setting is PAP or CHAP).

  When the Encapsulation setting is MPP, both the dialing side and the answering side of the link must support MP+. If only one side supports MP+, the connection uses MP or standard single-channel PPP. When the Encapsulation setting is PPP, the connection uses only PPP.

Usage: Enter a password. This password must match the one set up for your Pipeline in the RADIUS users file on the NAS (Network Access Server).

Dependencies: MPP Password applies only to outgoing MP+ calls in which the Authentication setting is PAP-Token-CHAP.

However, if the connection is already online over nailed-up channels, which end should dial to add switched channels? The only way to determine who calls (and therefore who is billed for the call) is by using this setting.

Usage:

• Check the box to specify that the Pipeline initiates the call.

  If it is checked, the Pipeline dials to bring online any switched circuits that are part of the call.

• If the box is unchecked, the Pipeline waits for the remote end to initiate the call.

  Unchecked is the default.

• If Fractional T1 Caller is not checked on the remote device, check Fractional T1 Caller on the local Pipeline; by the same token, if Fractional T1 Caller is checked on the remote device, uncheck Fractional T1 Caller on the local Pipeline.

• The Fractional T1 Caller setting applies only when Call Type is Nailed/MPP.

Usage: If checked, the Pipeline expects the connection to terminate and result in a call-back from the far-end device. Check this when the remote device requires callback security. This prevents problems that arise when Caller ID is required for the device that is expected to callback. If a call fails for any reason, regardless of whether or not the called machine requires Caller ID and is attempting a callback, the call initiator will still have to wait 90 seconds before attempting the call the same number again if this setting is checked.

If this setting is unchecked, the Pipeline does not expect call-back for this connection.

This is unchecked by default.

By default, any packet destined for the WAN causes the Pipeline to place a call. In addition, by default, every packet resets the idle timer, the indicator that the Pipeline uses to know when to clear a call. When you set up a Call Placement Filter, only those packets that the Call Placement Filter forwards can initiate a call or reset the Preempt or Idle Timeout settings.

Usage: Click the Call Placement Filter button. In the window that appears, select a filter and then click OK.

If you choose None, the Pipeline forwards all data packets. None is the default.

Dependencies: Keep this additional information in mind:

• If all channels of a link are nailed up (Call Type setting is Nailed in the Connection Profile), the Call Placement Filter setting does not apply (Call Placement Filter setting is dimmed).

• The Pipeline applies a Call Placement Filter after applying a data filter; only those packets that the data filter forwards can reach the Call Placement Filter.

• If IPX client bridging is in use (Discard RIP, SAP setting is Broadcasts), set the Call Placement Filter setting to None.

• Call Placement Filter in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Call Placement Filter setting takes precedence.

• If the Require Connection Profile setting is checked in the Answer tab, Call Placement Filter does not apply (Call Placement Filter setting is dimmed) in the Answer tab.

Frame Relay is an HDLC-based packet protocol that enables you to send data to a destination using one or more frame relay switches within a private network or a public carrier's network. HDLC stands for High Level Data Link Control.

From the viewpoint of the Pipeline, a frame relay switch is an endpoint for all DLCIs (Data Link Connection Indicators) connecting to it. A DLCI identifies a Connection Profile as a logical link. The frame relay switch connects the endpoints of the DLCIs to each other to make a virtual permanent circuit to which users can connect. The circuit acts like a wire between two endpoints with a fixed maximum bandwidth.

Usage: Enter a number between 16 and 991. The default is 16. Ask your frame relay network administrator for the value you should enter.

Dependencies: Keep this additional information in mind:

• Frame Relay DLCI appears in a Connection Profile only when the Encapsulation setting is Frame Relay.

• Each Connection Profile in which the Encapsulation setting is Frame Relay represents a separate logical link; you must assign it a unique setting for Frame Relay DLCI.

Usage: Enter the IP address of a remote station or router.

An IP address consists of four numbers between 0 and 255, separated by periods.

The default setting is 0.0.0.0; an answering Connection Profile with this setting matches all incoming IP addresses.

Example: 200.207.23.101

Dependencies: Keep this additional information in mind:

• The IP Address setting in the first Connection Profile is the same as the IP Address setting in the Configure Profile.

• The value of the IP Address setting on the local Pipeline must match the IP Address setting of the unit at the remote end of the link.

• No two calling Connection Profiles should have the same IP Address.

• Setting IP Address to 0.0.0.0 and clearing the Station setting resets all settings in the Connection Profile to their defaults.

• The IP Address setting does not apply (IP Address setting is dimmed) if the Pipeline does not support IP (Route IP setting is unchecked).

• If you do not know the right IP address to enter, you must obtain it from the network administrator.

  Do not attempt to configure an IP address by guesswork!

Usage:

• Check this box to disable advertising.
  The Pipeline does not advertise the IP address in RIP updates that it sends.

• Leave this unchecked to enable advertising.
  The Pipeline advertises the IP address in RIP updates that it sends.

  This is unchecked by default.

The Private Address setting does not apply (Private Address setting is dimmed) if the Pipeline does not support IP (Route IP setting is unchecked).

Usage: Click the up arrow to add bits to the subnet mask. Click the down arrow to take away bits from the subnet mask. The mask and the number of bits in the mask are shown to the right of the arrows.

Dependencies: Keep this additional information in mind:

• The value of the Subnet Mask setting must match the Subnet Mask setting expected by the device to which the Pipeline connects.

• If you do not know the right subnet mask to enter, you must obtain it from the network administrator. Do not attempt to configure the subnet mask by guesswork!

The WAN Alias setting applies only if the remote end of a link uses an implementation of PPP that requires that both ends of a WAN connection be on the same subnet.

If a router requires an IP number for each interface over which it sends or receives packets, that router is said to use numbered interfaces. The WAN Alias setting assigns a single IP number to all WAN lines connected to the Pipeline. Furthermore, the Pipeline assumes that all devices using numbered interfaces have agreed on the network number of the WAN; that is, if 10.0.2.1 is the Pipeline interface to the WAN, then the WAN has a network number 10.0.2.0 and all other devices using numbered interfaces agree to have a 10.0.2.x address.

Usage: Enter the IP address of the remote device.

An IP address consists of four numbers between 0 and 255, separated by periods. If a netmask is in use on the network, you must specify it. Separate the netmask from the IP address with a slash.

The default is 0.0.0.0/0.

Example: 200.207.23.7/24

Dependencies: The WAN Alias setting does not apply if the Pipeline does not support IP (Route IP setting is unchecked).

Usage: Enter the IP address of the numbered interface.

An IP address consists of four numbers between 0 and 255, separated by periods.

Example: 200.207.23.7/24

Dependencies: The Interface Address setting does not apply if the Pipeline does not support IP (Route IP is unchecked).

Usage: Click the up arrow to add bits to the subnet mask. Click the down arrow to take away bits from the subnet mask. The mask and the number of bits in the mask are shown to the right of the arrows.

Dependencies: Keep this additional information in mind:

• The value of the Subnet Mask setting must match the Subnet Mask setting expected by the device to which the Pipeline connects.

• If you do not know the right subnet mask to enter, you must obtain it from the network administrator. Do not attempt to configure the subnet mask by guesswork!

Usage: If checked, header compression is enabled.

For compression to occur, the Ascend unit must include the optional compression module. Dependencies: Keep this additional information in mind:

• This setting applies only to packets in TCP applications, such as Telnet.

  Telnet is a protocol used to link two computers in order to provide a terminal with a connection to the remote machine. The remote machine is known as the Telnet host. When you start a Telnet session, you connect to the Telnet host and log in. The connection enables you to work with the remote machine as though you were at a terminal connected to it.

• Turning on header compression is most effective in reducing overhead when the data portion of the packet is small.

Usage: Enter the IP address of a DNS server to be used for all connections that do not have a DNS server defined. The default value is 0.0.0.0.

Usage: Enter the IP address of a secondary DNS server to be used for all connections that do not have a DNS server defined. The default value is 0.0.0.0.

Usage: Check this to use client DNS servers. The default is unchecked.

Usage: Enter the IP address of the next hop router in dotted decimal notation. The default value is 0.0.0.0; if you accept this value, the Pipeline routes packets as specified in the routing table, using the system-wide default route if it cannot find a more specific route.

The Pipeline must have a direct route to the address you specify. The direct route can take place via a profile or an Ethernet connection. If the Pipeline does not have a direct route, it drops the packets on the connection. When you diagnose routing problems with a profile using this feature, an error in a per-user gateway address is not apparent from inspection of the global routing table.

Example: If you enter 10.0.0.3 for the Gateway setting, IP packets from the user with destinations through the default route will be routed through the gateway at 10.0.0.3.

If there are two routes available to a single destination network, you can ensure that the Pipeline uses any available nailed-up channel before using a switched channel by setting the Virtual Hops setting to a value higher than the metric of any nailed-up route. The higher the value entered, the less likely that the Pipeline will bring the link or route online. The Pipeline uses the lowest metric.

Usage: Enter a number between 1 and 15. This value is the virtual hop count. The default setting is 7.

Example: If a route to a station takes three hops over nailed-up lines, and the Virtual Hops setting is 4 in a Connection Profile that reaches the same station, the Pipeline does not bring the Connection Profile's link online.

Dependencies: Keep this additional information in mind:

• The Virtual Hops setting in a Connection Profile does not apply to bridged connections.

• If you enable RIP (Routing Information Protocol) across the WAN in a Connection Profile or an Answer tab, the hop count for the route can differ from the value of the Virtual Hops setting in the Route Profile because the Pipeline always uses the lower hop count.

• The hop count includes the metric of each switched link in the route.

• Virtual Hops only applies to IP connections.

When choosing which routes to put in the routing table, the router first compares the Priority values, preferring the lower number. If the Priority values are equal, the router compares the Virtual Hops setting values, using the route with the lower number of hops.

Usage: Enter a number between 0 and 255. The value of 255 means "Don't use this route"; this value is meaningful only for Connection Profiles.

Dependencies: These are the default values for different types of routes:

• The Routes learned from OSPF setting is 10.

• The Routes learned from ICMP Redirects setting is 30.

• The Routes learned from RIP setting is 100.

Note: Ascend recommends that all routers and hosts run RIP-v2 instead of RIP-v1. The IETF has voted to move RIP version 1 into the "historic" category and its use is no longer recommended.

Usage: Choose one of the following:

• Off
  This setting specifies that the Pipeline does not transmit or receive RIP updates.

  Off is the default.

• Send Only
  This setting specifies that the Pipeline transmits RIP updates, but does not receive RIP updates on this interface.

• Receive Only
  This setting specifies that the Pipeline receives RIP updates, but does not transmit RIP updates on this interface.

• Both
  This setting means that the Pipeline transmits and receives RIP updates on this interface.

The most significant difference between RIP (Routing Information Protocol) versions 1 and 2 is that RIP-v2 allows neighboring hosts to communicate netmasks to each other. RIP-v1 forces routers to guess the netmask.

If the Pipeline is communicating with other RIP-v2 routers and hosts, all routing tables contain the same addresses and routes. However, if the Pipeline is communicating with a RIP-v1 router, that router ignores the netmask field in the RIP-v2 packet, making use only of the IP address without the netmask. For this reason, we do not recommend that you run RIP-v1 and RIP-v2 on the same network in such a way that both RIP-v1 and RIP-v2 hosts hear each other's advertisements.

Note: Ascend recommends that all routers and hosts run RIP-v2 instead of RIP-v1. The IETF has voted to move RIP version 1 into the "historic" category and its use is no longer recommended.

Usage: Choose one of the following:

• 1 specifies to use RIP version 1

• 2 specifies to use RIP version 2

  2 is the default.

• The RIP Version setting does not apply if the Pipeline does not support IP (Route IP setting is unchecked).

• RIP Version -and- RIP Direction in the Answer tab apply to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its RIP Version -and- RIP Direction setting takes precedence.

• If the Require Connection Profile setting is checked in the Answer tab, RIP Version -and- RIP Direction does not apply (RIP Version and RIP Direction settings are dimmed) in the Answer tab.

Usage: Choose one of the following:

• Router
  This setting specifies that the caller is an IPX router.

  Router is the default.

• Dialin
  This setting specifies that the caller is a dial-in NetWare client that incorporates PPP software and dial-out hardware, but does not have an Ethernet interface.

  This setting causes the Pipeline to assign the caller an IPX address derived from the value of WAN Network Number.

Usage:

• If the checkbox is checked, the Pipeline places a call to the location specified in the Connection Profile when a workstation looks for the nearest server.

  Note that a workstation is likely to stop attempting to find a server before the Pipeline establishes any connections with the Dial on Query mechanism.

• If the checkbox is not checked, the Pipeline does not place a call to the location specified in the Connection Profile when a workstation looks for the nearest server.

  This is unchecked by default.

When a Pipeline is used to connect NetWare clients to a very large IPX network, the IPX routing table created by the Pipeline may become very large and unmanageable, and can cause the Pipeline to run out of memory. As an alternative to maintaining these large routing tables locally, the Pipeline may have a static IPX route to the corporate network and disable IPX RIP. Either end of the WAN link may disable or fine-tune IPX RIP behavior.

Usage: Choose one of the following:

• Off
  This indicates that the device will neither send nor receive IPX RIP updates on this WAN link.

• Send Only
  This indicates that the device will send RIP updates but will not receive them.

• Receive Only
  This indicates that the device will receive RIP updates but will not send them.

• Both
  This indicates that the device will both send and receive RIP updates on this WAN link.

  Both is the default.

When a Pipeline is used to connect NetWare clients to a very large IPX network, the IPX service table created by the Pipeline may become very large and unmanageable, and can cause the Pipeline to run out of memory. As an alternative to maintaining these large service tables locally, the Pipeline may create static service table entries and turn off IPX SAP. Either end of the WAN link may disable or fine-tune IPX SAP behavior.

Usage: Choose one of the following:

• Off
  This indicates that the device will neither send nor receive SAP Direction updates on this WAN link.

• Send Only
  This indicates that the device will send SAP updates but will not receive them.

• Receive Only
  This indicates that the device will receive SAP updates but will not send them.

• Both
  This indicates that the device will both send and receive SAP updates on this WAN link.

  Both is the default.

The value of IPX Net # specifies the network number of the router at the remote end of the connection.

Usage: Enter an Ethernet network number using an 8-digit (4-byte) hexadecimal value. The default is 00000000.

Specify the network number of the router at the remote end of the connection only if the router requires that the Pipeline know its network number before connecting. You almost never need to set this setting in a Connection Profile.

If you accept the default of 00000000, the Connection Profile is still valid, but the Pipeline does not advertise the route until it makes a connection to the Ethernet network.

Example: DE040600

Dependencies: The IPX Net # setting does not apply (IPX Net # setting is dimmed) if the Pipeline is not set up for IPX routing (Route IPX setting is unchecked).

Generally, you need to enter a value in this setting only if the Pipeline operates with a non-Ascend router that uses a numbered interface. It does not apply if you are routing from one Pipeline to another, or to a router that does not use a numbered interface.

Usage: Enter an appropriate network number. The default value is 00000000. FFFFFFFF is invalid.

Dependencies: For the IPX Alias # setting to apply, you must enable IPX routing in the Connection Profile by checking Route IPX.

Usage: Choose one of the following:

• None
  This setting specifies that special IPX behavior does not take place.

  Choose this setting when the LAN on each side of the bridge has one or more IPX servers.

  None is the default.

• Broadcasts
  This setting specifies that the Pipeline discards RIP (Routing Information Protocol) and SAP (Service Advertising Protocol) periodic broadcasts at its WAN interface, but forwards RIP and SAP queries.

  The WAN interface is the port on the Pipeline that is connected to a WAN line. RIP and SAP queries enable a client workstation to locate a NetWare server across the network. Choose this setting when both these conditions are true:

  • The local LAN has IPX clients but no servers.

  • The Pipeline is acting as a bridge to another LAN containing only IPX servers or a combination of IPX servers and clients.

  
  
• Broadcasts & Queries
  This setting specifies that the Pipeline discards all RIP (Routing Information Protocol) and SAP (Service Advertising Protocol) periodic broadcasts and queries at its WAN interface.

  Server mode allows the Pipeline to bring down calls during idle periods without breaking client-server or peer-to-peer connections.

  Ordinarily, when a NetWare server does not receive a reply to the watchdog session keepalive packets it sends to a client, it closes the connection. When you select Server mode, however, the Pipeline replies to NCP watchdog requests on behalf of clients on the other side of the bridge; in other words, the Pipeline tricks the server watchdog process into believing that the link is still active. This process is called watchdog spoofing.

  Choose this setting when both these conditions are true:

  • The Pipeline is acting as a bridge to a remote LAN with IPX clients, but no servers.

  • The local LAN contains only IPX servers, or a combination of IPX clients and servers.

• If you select the Server setting, you must also specify a value for the Watchdog Spoof Time setting, indicating the maximum length of idle time during which the Pipeline performs watchdog spoofing for NetWare connections.

• If the connection does not bridge (Enable Bridging setting is unchecked), the Discard RIP, SAP setting does not apply (Discard RIP, SAP setting is dimmed).

• If the encapsulation for the connection is Frame Relay (Encapsulation setting is Frame Relay), the Discard RIP, SAP setting does not apply (Discard RIP, SAP setting is dimmed).

• If you have not specified an IPX frame type (Frame Type setting is None), the Discard RIP, SAP setting does not apply (Discard RIP, SAP setting is dimmed).

• We highly recommend that you check the Dial on Broadcast setting when Discard RIP, SAP setting is Client and uncheck it when Discard RIP, SAP setting is Server.

  When a client on the local Ethernet interface sends out broadcast packets to locate a server, and the Pipeline must bridge these packets to another network, the Pipeline starts up a session for each Connection Profile in which Dial on Broadcast setting is checked. The server need not broadcast and then dial, so set Dial on Broadcast setting is unchecked to keep broadcast packets from causing the Pipeline to dial automatically.

• If the Pipeline on one LAN sets Discard RIP, SAP setting to Server and the LAN on the other side of the connection has only NetWare clients, the Pipeline on the client-only LAN should set Discard RIP, SAP setting to Client.

  If both LANs contain servers, both sides of the connection should set Discard RIP, SAP setting to None.

• Although Discard RIP, SAP setting is dimmed if Enable Bridging setting is unchecked or Frame Type setting is None, the Pipeline automatically performs watchdog spoofing just as though you had set Discard RIP, SAP setting to Server; however, the Pipeline does not filter as though you had set Discard RIP, SAP setting to Server.

Ordinarily, when a NetWare server does not receive a reply to the watchdog session keepalive packets it sends to a client, it closes the connection. When you select Server mode for the Discard RIP, SAP setting, however, the Pipeline replies to NCP watchdog requests on behalf of clients on the other side of the bridge; in other words, the Pipeline tricks the server watchdog process into believing that the link is still active.

The time period for watchdog spoofing specified by the Watchdog Spoof Time setting begins when the WAN session goes offline. If the WAN session reconnects, the Pipeline cancels the timeout.

Watchdog Spoof Time applies when the Pipeline is on a LAN containing a NetWare server.

Usage: Enter the timeout value in minutes. You can enter any value from 0 to 65535. If you enter 0, the Pipeline responds to server watchdog requests indefinitely. 30 is the default.

Dependencies: The Watchdog Spoof Time setting does not apply (Watchdog Spoof Time setting is dimmed) if Discard RIP, SAP setting is None.

• IPX SAP Input filters apply to all SAP packets that the Ascend unit receives.

  Input filters screen advertised services and exclude them from its service table as specified in the filters.

• IPX SAP Output filters apply to SAP response packets that the Ascend unit transmits.

  If the Ascend unit receives a SAP request packet, it applies Output filters before transmitting the SAP response, and excludes services from the response packet as specified in the filters.

(To define filters, go to the Filters tab.)

Usage: Choose one of the following:

• Don't Monitor
  This specifies that the Ascend unit will not monitor traffic over the link and will not use DBA.

• Transmit
  This specifies that the Ascend unit will add or subtract bandwidth based on the amount of data it transmits.

• Transmit & Receive
  This specifies that the Ascend unit will add or subtract bandwidth based on the amount of data it transmits or receives.

  Transmit & Receive is the default.

The Pipeline uses the historical time period specified by the Sample Time setting as the basis for calculating ALU. It then compares ALU to the amount specified in the Target % setting. When ALU exceeds the threshold defined by Target % for a period of time greater than the value of the Add Time setting, the Pipeline attempts to add a channel. When ALU falls below the threshold defined by Target % for a period of time greater than the value of the Sub Time setting, the Pipeline attempts to remove a channel.

MP+ supports Dynamic Bandwidth Allocation.

Usage: Choose one of the following:

• Constant
  This gives equal weight to all samples taken during the historical time period specified by the Sample Time setting.

  When you select this, older historical samples have as much impact on the decision to change bandwidth allocation as do more recent samples.

• Linear
  This gives more weight to recent samples of bandwidth usage than to older samples taken during the historical period specified by the Sample Time setting; the weighting grows at a linear rate.

• Quadratic
  This gives more weight to recent samples of bandwidth usage than to older samples taken during the historical period specified by the Sample Time setting; the weighting grows at a quadratic rate.

  Quadratic is the default for MP+ calls (Encapsulation setting is MPP).

Dependencies: Keep this additional information in mind:

• To dynamically allocate bandwidth by tracking line usage, you use the Add Time, DBA Algorithm, Max Channels, Min Channels, Sample Time, Sub Time, and Target % settings.

• DBA Algorithm in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, its DBA Algorithm setting takes precedence.

When ALU exceeds the Target % threshold for a period of time greater than the value of the Add Time setting, the Pipeline attempts to add a channel.When ALU falls below the Target % threshold for a period of time greater than the value of the Sub Time setting, the Pipeline attempts to remove a channel.

The number of seconds you choose for the Sample Time setting depends on your device's traffic patterns. For example, if you want to average spikes with normal traffic flow, you may want the Pipeline to establish a longer historical time period. If, on the other hand, traffic patterns consist of many spikes that are short in duration, you may want to specify a shorter period of time; doing so assigns less weight to the short spikes.

Usage: Enter a number between 1 and 300. The default value is 15 seconds.

Dependencies: Keep this additional information in mind:

• The Sample Time setting applies only to dynamic MP+ calls (Encapsulation setting is MPP).

• If you specify a small value for the Sample Time setting, and increase the values of the Add Time setting and the Sub Time setting relative to the value of Sample Time, the system becomes less responsive to quick spikes.

• The easiest way to determine the proper values for Sample Time, Add Time, and Sub Time is to observe usage patterns; if the system is not responsive enough, the value of Sample Time is too high.

Usage: Enter a number between 1 and the maximum number of channels your system supports. The default setting is 1.

The maximum value of the Base Channels setting depends on the encapsulation method that both ends of the link use.

• For an PPP link (for which Encapsulation setting is PPP), the value of Base Channels is always 1.

• For an MP+ or MP link (for which Encapsulation setting is MPP), the amount you specify is limited by the number of channels available, but the device at the remote end of the link must also support MP+ or MP.

Dependencies: Keep this additional information in mind:

• You can determine the base bandwidth of a call by multiplying the value of the Base Channels setting by the value of the Link Type setting.

• The Base Channels setting does not apply (Base Channels setting is dimmed) when all channels of the link are nailed up (Call Type setting is Nailed).

• For optimum MP+ performance, both sides of a connection must set these settings to the same values:
  • Base Channels

  • Min Channels

  • Max Channels

Usage: Enter a number between 1 and the maximum number of channels your system supports. The default setting is 1.

Dependencies: The Min Channels setting applies only to MP+ calls (Encapsulation setting is MPP). For optimum MP+ performance, both sides of a connection must set these settings to the same values:

• Base Channels

• Min Channels

• Max Channels

Usage: Enter a number between 1 and the maximum number of channels your system supports.

Dependencies: Keep this additional information in mind:

• The Max Channels setting applies only to dynamic MP+ calls (Encapsulation setting is MPP).

• For optimum MP+ performance, both sides of a connection must set these settings to the same values:
  • Base Channels

  • Min Channels

  • Max Channels

• Max Channels in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, its Max Channels setting takes precedence.

When utilization exceeds the threshold for a period of time greater than the value of the Add Time setting, the Pipeline attempts to add a channel. Using the Add Time and Sub Time settings prevents the system from continually adding and subtracting bandwidth, and can slow down the process of allocating or removing bandwidth.

Usage: Enter a number between 1 and 300.

When the Pipeline is using MP+ (Encapsulation setting is MPP), the default value is 5.

Dependencies: Keep this additional information in mind:

• Additional channels must be available, and the number of channels added cannot exceed the amount specified by the Max Channels setting.

• Add Time in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Add Time setting takes precedence.

• If Require Connection Profile setting is checked in the Answer tab, Add Time does not apply (Add Time setting is dimmed) in the Answer tab.

• Add Time and Sub Time have little or no effect on a system with a high Sample Time value.

  If the value of Sample Time is low, the Add Time and Sub Time settings provide an alternative way to ensure that spikes must persist for a certain period of time before the system responds.

When utilization falls below the threshold for a period of time greater than the value of the Sub Time setting, the Pipeline attempts to remove a channel. Using the Add Time and Sub Time settings prevents the system from continually adding and subtracting bandwidth, and can slow down the process of allocating or removing bandwidth.

Usage: Enter a number between 1 and 300.

When the Pipeline is using MP+ (Encapsulation setting is MPP), the default value is 10.

Dependencies: Keep this additional information in mind:

• One channel must be up at all times.

• Removing bandwidth cannot (a) cause the ALU to exceed the threshold specified by the Target % setting or (b) cause the number of channels to fall below the amount specified by the Min Channels setting.

• Sub Time in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Sub Time setting takes precedence.

• Add Time and Sub Time have little or no effect on a system with a high Sample Time value.

  However, if the value of Sample Time is low, the Add Time and Sub Time settings provide an alternative way to ensure that spikes persist for a certain period of time before the system responds.

Usage: Enter a number between 0 and 65535. If you specify 0 (zero), the Pipeline does not enforce a limit; an idle connection stays open indefinitely.

This setting specifies the target percentage of bandwidth utilization for an MP+ call (Encapsulation setting is MPP).

The Pipeline uses the historical time period specified by the Sample Time setting as the basis for calculating average line utilization (ALU) of transmitted data. It then compares ALU to the amount specified in the Target % setting.

When ALU exceeds the threshold defined by Target % for a period of time greater than the value of the Add Time setting, the Pipeline attempts to add a channel. When ALU falls below the threshold defined by Target % for a period of time greater than the value of the Sub Time setting, the Pipeline attempts to remove a channel.

Usage: Enter a number between 0 and 100.

The default is 70. When the value is 70%, the device adds bandwidth when it exceeds a 70 percent utilization rate, and subtracts bandwidth when it falls below that number.

Dependencies: When selecting a target utilization value, keep these guidelines in mind:

• Monitor how the application behaves when using different bandwidths.

  For example, an application might be able to use 88% of a 64-Kbps link, but only 70% of a 256-Kbps link.

• Monitor the application at different loads.

Usage: Enter a number between 0 and 99. The default value is 0; this setting causes the Pipeline to ignore bandwidth utilization when determining whether to clear a call.

Dependencies: Keep this additional information in mind:

• MP+ must be the selected encapsulation method (Encapsulation setting is MPP) in a Connection Profile.

• If the device at the remote end of the link enters an Idle % setting lower than the value you specify, the Pipeline does not clear the call until bandwidth utilization falls below the lower percentage.

• If either end of a connection sets the Idle % setting to 0 (zero), the Pipeline ignores bandwidth utilization when determining when to clear a call.

• If the time set by the Idle Timeout setting expires, the call disconnects whether or not bandwidth utilization falls below the Idle % setting.

• When bandwidth utilization falls below the Idle % setting, the call disconnects regardless of whether the time specified by the Idle Timeout setting has expired.

• Because the Idle % setting is dependent on traffic levels on both sides of the connection, we recommend that you use the Idle Timeout setting in preference to it.

• Idle % in the Answer tab applies to incoming calls for which no Connection Profile exists; if a Connection Profile exists, the setting of its Idle % setting takes precedence.

• If Require Connection Profile setting is checked in the Answer tab, Idle % does not apply (Idle Timeout setting is dimmed) in the Answer tab.

Usage: Check the box to make the filter persistent.

• If checked, the filter or firewall assigned to this Connection Profile will persist after the connection has been torn down.

  Note: Typically a firewall will persist for about an hour after its associated connection has been torn down.

If the primary connection is unavailable, the Pipeline automatically diverts traffic to the backup connection. A connection can fail if, for example, a frame relay connection loses a Permanent Virtual Circuit, the physical link fails, or if a T1 line is in a red alarm condition. When the primary connection is restored, traffic again uses the primary connection.

When you use the backup connection, the Pipeline does not move routes to the backup profile. Therefore, the IP routes shown in the terminal server display may be incorrect, although statistical counts reflect the change.

Usage: Click the Backup Connection button, choose a connection, and then click OK. The backup connection can be switched or nailed up.

Dependencies: Keep this additional information in mind:

• Do not create nested backup connections.

• The Backup Connection setting applies only to nailed-up connections (for which Call Type setting is Nailed or Nailed/MPP); otherwise, Backup Connection setting is dimmed.

• Settings that you define in the primary Connection Profile do not automatically apply to the backup Connection Profile.

  For example, if you set the primary Connection Profile to filter Telnet packets, you must set the backup profile to filter Telnet packets as well. Outgoing Frame Relay packets are the only packets that follow the primary Connection Profile definitions. All other packets follow the backup Connection Profile definitions.

• Backup Connection is intended for situations in which the remote device (such as a data center) goes out of service; the backup call is made to a backup data center.

  Backup Connection is not intended to provide alternative lines for getting to a single destination.

• Do not confuse the Backup Connection setting with the Secondary Connection setting. A backup Connection Profile is used to re-establish an existing connection that has terminated; a Secondary Connection Profile is used to establish a new connection if the primary Connection Profile cannot. That is, the Secondary Connection provides an alternative line for a single destination, which the backup Connection Profile does not.

Usage: Click the Secondary Connection button, choose a connection, and then click OK. The backup connection can be switched or nailed up.

Dependencies: Keep this additional information in mind:

• Secondary Connection Profiles cannot be nested. That is, secondary Connection Profiles cannot also have Secondary Connection profiles.

• Do not confuse the Secondary Connection setting with the Backup Connection setting. A Backup Connection setting is used to re-establish an existing connection that has terminated; a Secondary Connection setting is used to establish a new connection if the primary Connection Profile cannot.

• Settings that you define in the primary Connection Profile do not automatically apply to the secondary Connection Profile.

  For example, if you set the primary Connection Profile to filter Telnet packets, you must set the secondary profile to filter Telnet packets as well.

• Outgoing Frame Relay packets are the only packets that follow the primary Connection Profile definitions. All other packets follow the backup Connection Profile definitions.