Traffic Classification Rules
Traffic Classification rules allow you to assign VLAN membership and/or class of service to your network traffic based on the traffic's classification type. Classification types are derived from Layers 2, 3, and 4 of the OSI model, and all network traffic can be classified according to specific layer 2/3/4 information contained in each frame. In Policy Manager, rules are used to provide four key policy features: traffic containment, traffic filtering, traffic security, and traffic prioritization. Examples of how to design rules for each of these features are given below.
A Traffic Classification rule has two main parts: Traffic Description and Actions. The Traffic Description identifies the traffic classification type for the rule. The Actions specify whether traffic matching that classification type will be assigned VLAN membership, class of service, or both. When a frame arrives on a port, the switch checks to see if the frame's classification type matches the type specified in a rule. If it does, then the actions defined in that rule will apply to the frame. Use the Policy Manager's Rule Wizard to quickly and easily create a rule and define its traffic description and actions.
In Policy Manager, rules are created and then grouped together into Services, which are then used to define roles. A role is assigned to each port either through end user authentication or as the port's default role. This means that there can be multiple rules active on a port. When a frame is received on a port, if the frame's classification type matches more than one rule, classification precedence rules are used to determine which rule to use.
The following information is discussed in this file:
- Traffic Descriptions
- Actions
- Classification Types and their Parameters
- Examples of How Rules are Used
- Classification Precedence Rules
Traffic Descriptions
When you create a Traffic Classification rule in Policy Manager, you must define the rule's traffic description. The traffic description identifies the traffic classification type for that rule. You must select a classification type, and then select or enter certain parameters or values for each type.Classification types are grouped according to Layers 2, 3, and 4 of the OSI model and there are multiple classification types for each layer.
OSI Model |
---|
Layer 7 - Application |
Layer 6 - Presentation |
Layer 5 - Session |
Layer 4 - Transport |
Layer 3 - Network |
Layer 2 - Data Link |
Layer 1 - Physical |
Specific Layer 2/3/4 information contained in the header of each frame is used to identify the frame's classification type. Each layer uses different information to classify frames.
- Layer 2 Data Link -- classifies frames based on an exact match of the MAC address or specific protocol type of each frame.
- Layer 3 Network -- classifies IP or IPX frames based on specific information contained within the Layer 3 header.
- Layer 4 Transport -- classifies IP frames based on specific Layer 4 TCP or UDP port numbers contained in the header.
Actions
When you create a Traffic Classification rule in Policy Manager, you must define the actions that the rule will perform. When a frame arrives on a port, the switch checks to see if the frame's classification type matches the type specified in a rule. If it does, then the actions defined in that rule will apply to the frame. Actions specify whether the frame will be assigned VLAN membership, class of service, or both.VLAN Membership
In your network domains, you can create VLANs (Virtual Local Area Networks) that allow end systems connected to separate ports to send and receive traffic as though they were all connected to the same network segment. Using traffic classification rules, you can classify a frame based on the frame's classification type to have membership in a specific VLAN, providing important traffic containment, filtering, and security for your network.For example, a network administrator could use rules to separate end-user traffic into VLANs according to protocol, subnet, or application. Rules could also be used to group geographically separate end systems into job specific workgroups.
Priority (Class of Service)
Traffic Classification rules allow you to assign a transmission priority to frames received on a port based on the frame's classification type. For example, a network administrator could use rules to assign priority to one network application over another.Priority is a value between 0 and 7 assigned to each frame as it is received on a port, with 7 being the highest priority. Frames assigned a higher priority will be transmitted before frames with a lower priority. Each of the priorities is mapped into a specific traffic queue by the switch or router. The insertion of the priority value (0-7) allows all 802.1Q devices in the network to make intelligent forwarding decisions based on its own level of support for prioritization.
The Matrix product line supports four (0-3)or eight (0-7) traffic queues per port. The number of traffic queues varies by port type. These queues can be serviced based on a strict method, meaning that all frames in Queue 3 will be transmitted before the frames in Queue 0, or based on a fair weighted method. The weighted method allows the network administrator to give a certain percentage or weight to each queue, preventing a lower priority queue from being starved.
Policy Manager enables you to utilize priority by creating classes of service that include an 802.1p priority (rate-limited or not) and/or IP type of service values. You can then assign the class of service as a classification rule action, as part of the definition of an automated service, or as a role default. See How to Define a Bandwidth Rate Limiter and How to Create a Class of Service for more information.
Classification Types and their Parameters
When you define a rule's traffic description, you select a classification type, and then select or enter certain parameters or values for each type. Classification types are grouped according to Layers 2, 3, and 4 of the OSI model and there are multiple classification types for each layer.Layer 2 -- Data Link Classification Types
Layer 2 classification types allow you to define classification rules based on an exact match of the MAC address or specific protocol type of each frame.
-
Ethertype
-
This classification type is based on the specific protocol type of each frame defined in the two-byte Ethertype field. Each protocol has a corresponding classification type address that is automatically entered into the
Value field. If you select
Other, you must manually enter the value in hexadecimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices. Common Ethertypes and their values:
IP 0x0800 ARP 0x0806 Reverse ARP 0x8035 Novell IPX 1 0x8137 Novell IPX 2 0x8138 Banyan 0x0bad AppleTalk 0x809b AppleTalk ARP 0x80f3 DECnet Phase 4 0x6003 DEC MOP Dump/Load 0x6001 DEC MOP Remote Console 0x6002 DEC LAT 0x6004 DEC Diagnostic Protocol 0x6005 DEC Customer Protocol 0x6006 DEC LAVC, SCA 0x6007 DEC Unassigned 0x6008-0x6009 Other 0x0600-0xFFFF
-
DSAP/SSAP
-
This classification type is based on the specific protocol type of each frame defined in the DSAP and SSAP fields. Each protocol has a corresponding classification type address that is automatically entered into the
Value field. If you select
Other, you must manually enter the value in hexadecimal form. You can also enter advanced configurations, however, these are only supported on Matrix N-Series Platinum devices. The DSAP and SSAP values are one byte each, and these values must match. They are combined to create one two-byte value (e.g., a DSAP value of e0 and an SSAP value of e0 are combined to create a two-byte value of e0e0). Common DSAP/SSAP types and their values:
IP 0x0606 IPX 0xe0e0 NetBIOS 0xf0f0 Banyan Vines 0xbcbc SNA 0x0404 Other a two-byte value
-
MAC Address Source, MAC Address Destination, MAC Address Bilateral
- These classification types are based on an exact match of the source, destination, or bilateral (either source or destination) MAC address contained in an Ethernet frame. You can select a mask, however, masking a MAC address is only supported on Matrix N-Series Platinum devices.
Layer 3 -- Network Classification Types
Layer 3 Network classification types allow you to define classification rules based on specific information contained within the Layer 3 header of an IP or IPX frame.-
IP Type of Service
-
This classification type is based on an exact match of the one-byte TOS/DSCP field contained in the IP header of a frame. The TOS (Type of Service) or DSCP (Diffserve Codepoint) value is defined by an 8-bit hexadecimal number between 0 and FF.
Type of Service can be used by applications to indicate priority and Quality of Service for each frame. The level of service is determined by a set of service parameters which provide a three way trade-off between low-delay, high-reliability, and high-throughput. The use of service parameters may increase the cost of service. In many networks, better performance for one of these parameters is coupled with worse performance on another. Except for very unusual cases, at most, two of the parameters should be set.
For a TOS value, the 8-bit hexadecimal number breaks down as follows:
Bits 0-2: Precedence Bit 3: 0=Normal Delay, 1=Low Delay Bit 4: 0=Normal Throughput, 1=High Throughput Bit 5: 0=Normal Reliability, 1=High Reliability Bits 6-7: Explicit Congestion Notification The precedence bits (bits 0-2) break down as follows:
111 - Network Control 110 - Internetwork Control 101 - CRITIC/ECP 100 - Flash Override 011 - Flash 010 - Immediate 001 - Priority 000 - Routine The Network Control precedence designation is intended to be used within a network only. The actual use and control of that designation is up to each network. The Internetwork Control designation is intended for use by gateway originators only.
For a DSCP value, the value represents codepoints for two Differentiated Services (DS) Per-Hop-Behavior (PHB) groups called Expedited Forwarding (EF) and Assured Forwarding (AF). For more information on these PHB groups, refer to RFC 2597 and RFC 2598.
For information on how to automatically generate a TOS or DSCP value, see TOS/DSCP Configuration window.
-
IP Protocol Type
-
This classification type is based on the specific protocol type defined in a field contained in the IP header of each frame. Each protocol has a corresponding classification type address that is automatically entered into the
Value field. If you select
Other, you must manually enter the value in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices.
Common IP Protocol types and their values:TIP: You can define a new IP Protocol Type using the Pre-Defined Well-Known IDs window. Once defined, it is available for selection from the list of well-known values when defining the rule's traffic classification type. ICMP 1 IGMP 2 IPv6 41 L2TP 115 OSPF 89 PIM 103 RSVP 46 TCP 6 UDP 17 VRRP 112 Other 0-255
-
IP Address Source, IP Address Destination, IP Address Bilateral
- These classification types are based on an exact match of the source, destination, or bilateral (either source or destination) IP address information contained within the IP header of each frame. This classification type must also contain a Mask entry along with the specific IP address. This Mask allows a switch to determine whether a classification is based on a specific IP address, IP subnet, or range within an IP subnet. For example, a user could define a rule where all frames from the 129.168.28.x network are classified into the Red VLAN by setting the IP address of 129.168.28.0 with the mask of 255.255.255.0.
-
IP Fragment
- This classification type is based on Layer 4 information in fragmented frames. IP supports frame fragmentation, where large frames are divided into smaller fragments and sent wrapped in the original Layer 3 (IP) header. When a frame is fragmented, information that is Layer 4 and above is only present in the first fragment. For example, the first fragment may be classified to Layer 4, while subsequent fragments will be classified only to Layer 3. The Matrix product line does not support Layer 4 classification for IP frames that have been fragmented, as the Layer 4 information is not present in these frames. Using the IP Fragment classification rule, any frame which is a fragment of a larger frame, is classified according to the information in the original frame. If the first fragment is classified to Layer 4, subsequent fragments will also be classified to Layer 4.
-
IPX Class of Service
- This classification type is based on specific information contained within the Layer 3 header of an IPX frame. This is a one-byte field used for transmission control (hop count) by IPX routers. Enter a valid IPX Class of Service in decimal form, 0-255. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices.
-
IPX Packet Type
-
This classification type is based on specific information contained within the Layer 3 header of an IPX frame. Each IPX Packet type has a corresponding classification type address that is automatically entered into the
Value field. If you select
Other, you must manually enter the value in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices. Common IPX Packet types and their values:
Hello/SAP 0 RIP 1 Echo Packet 2 Error Packet 3 NetWare 386 4 SeqPackProt 5 NetWare 286 17 Other 0-31
-
IPX Network Source, IPX Network Destination, IPX Network Bilateral
- These classification types are based on specific information contained within the Layer 3 header of an IPX frame. It is a four-byte user-defined value that represents the IPX source, destination, or bilateral (either source or destination) network number. This value must be a valid IPX network address in hexadecimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices.
-
IPX Socket Source, IPX Socket Destination, IPX Socket Bilateral
-
These classification types are based on specific information contained within the Layer 3 header of an IPX frame. It is a two-byte, user-defined value that represents the IPX source, destination, or bilateral (either source or destination) socket numbers. This value is used by higher layer protocols to target specific applications running among hosts. Each IPX Socket type has a corresponding classification type address that is automatically entered into the
Value field. If you select
Other, you must manually enter the value in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices. Common IPX Socket types and their values:
NCP 1105 SAP 1106 RIP 1107 NetBIOS 1109 Diagnostics 1110 NSLP 36865 IPX Wan 56868 Other 0-65535
-
ICMP
-
This classification type is based on an exact match of the ICMP (Internet Control Message Protocol) message contained in the ICMP tag within a frame. Each ICMP message has a corresponding value that is automatically entered in the
Value field. If you select
Other, you must manually enter the value in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices. Common ICMP messages and their values:
Echo Reply 0 Destination Unreachable 3 Source Quench 4 Redirect 5 Echo 8 Router Advertisement 9 Router Solicitation 10 Time Exceeded 11 Parameter Problem 12 Timestamp 13 Timestamp Reply 14 Information Request 15 Information Reply 16 Address Mask Request 17 Address Mask Reply 18
-
VLAN ID
- This classification type is based on an exact match of the VLAN tag contained within a frame. Select a VLAN ID (VID) from the list of VLANs defined in Policy Manager. If you select Other, you must enter a single VID or specify a range of VIDs in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices.
-
Priority
- This classification type is based on an exact match of the Priority tag contained within a frame. Select a Priority value 0 - 7 from the list of well-known values, or select Other and enter a value in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices.
Layer 4 -- Application Transport Classification Types
Layer 4 IP classification types allow you to define classification rules based on specific Layer 4 TCP or UDP port numbers contained in the header of an IP frame. You can specify a specific port number or a range of port numbers.Note: The Matrix product line does not support Layer 4 classification for IP frames that have been fragmented, as the Layer 4 information is not present in these frames. If a Matrix device has an FDDI HSIM installed, Layer 4 classification will not be supported for any frames larger than 1500 bytes. Frames larger than 1500 bytes are fragmented internally in the switch. When creating classification rules based on specific Layer 4 information, using the IP Fragment classification rule will allow fragmented frames to be classified according to the Layer 4 information contained in the original frame.
-
IP UDP Port Source, IP UDP Port Destination, IP UDP Port Bilateral
-
These classification types are based on specific Layer 4 UDP port numbers contained within the header of an IP frame. Each UDP type has a corresponding classification type address that is automatically entered into the
Value field. If you select
Other, you must manually enter the value in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices. UDP port numbers are defined in RFC 1700.
Common UDP types and their values:TIP: You can define a new value for a UDP port number using the Pre-Defined Well-Known IDs window. Once defined, it is available for selection from the list of well-known values when defining the rule's traffic classification type. FTP Data 20 FTP 21 SSH 22 Telnet 23 SMTP 25 TACACS 49 DNS 53 Bootps 67 Bootpc 68 TFTP 69 Finger 79 HTTP 80 POP3 110 Portmapper 111 NNTP 119 NTP 123 NetBIOS Name Service 137 NetBIOS Datagram 138 NetBIOS Session Service 139 IMAP2 143 SNMP 161 IMAP3 220 LDAP 389 HTTPS 443 R-Exec 512 R-Login 513 R-Shell 514 LPR 515 SOCKS 1080 CITRIX ICA 1494 RADIUS 1812 RADIUS Accounting 1813 NFS 2049 Other 0-65535
-
IP TCP Port Source, IP TCP Port Destination, IP TCP Port Bilateral
-
These classification types are based on specific Layer 4 TCP port numbers contained within the header of an IP frame. Each TCP type has a corresponding classification type address that is automatically entered into the
Value field. If you select
Other, you must manually enter the value in decimal form. You can enter a range of values, however, range rules are only supported on Matrix N-Series Platinum devices. TCP port numbers are defined in RFC 1700.
Common TCP types and their values:TIP: You can define a new value for a TCP port number using the Pre-Defined Well-Known IDs window. Once defined, it is available for selection from the list of well-known values when defining the rule's traffic classification type. FTP Data 20 FTP 21 SSH 22 Telnet 23 SMTP 25 TACACS 49 DNS 53 Bootps 67 Bootpc 68 TFTP 69 Finger 79 HTTP 80 POP3 110 Portmapper 111 NNTP 119 NTP 123 NetBIOS Name Service 137 NetBIOS Datagram 138 NetBIOS Session Service 139 IMAP2 143 SNMP 161 IMAP3 220 LDAP 389 HTTPS 443 R-Exec 512 R-Login 513 R-Shell 514 LPR 515 SOCKS 1080 CITRIX ICA 1494 RADIUS 1812 RADIUS Accounting 1813 NFS 2049 Other 0-65535
-
IP UDP Port Source Range, IP UDP Port Destination Range, IP UDP Port Bilateral Range
- These classification types are based on Layer 4 UDP port numbers contained within the header of an IP frame. When you select this type, you enter a range of UDP port numbers that the port number in the header will be matched against. Enter the lower and upper range values in decimal form. UDP port numbers are defined in RFC 1700.
-
IP TCP Port Source Range, IP TCP Port Destination Range, IP TCP Port Bilateral Range
- These classification types are based on Layer 4 TCP port numbers contained within the header of an IP frame. When you select this type, you enter a range of TCP port numbers that the port number in the header will be matched against. Enter the lower and upper range values in decimal form. TCP port numbers are defined in RFC 1700.
Examples of How Rules are Used
Traffic Classification rules are used to provide four key policy features: Traffic Containment, Traffic Filtering, Traffic Security, and Traffic Priority.Traffic Containment
Using classification rules, network administrators can group together users of a given protocol, subnet, or application, and control where their traffic can logically go on the network.
The figure above shows a configuration where the network administrator wants to separate end-user traffic into VLANs based on the assigned IP subnet of each department. This can easily be accomplished by creating two Layer 3 classification rules based on the IP subnet range of the respective departments.
Rule 1 - Engineering, which uses the 132.181.28.x subnet, will be assigned to the Red VLAN.
Rule 2 - Sales, which uses the 132.181.29.x subnet, will be assigned to the Blue VLAN.
Based on these two Layer 3 classification rules, the traffic from the Engineering VLAN will be isolated from the Sales VLAN. Since these rules are based on Layer 3 information, an Engineering user could enter the network from a connection in the Sales department, and that user would still be contained in the Engineering VLAN.
Traffic Filtering
Classification rules can also be used to filter specific unwanted traffic. Filter criteria can include things such as broadcast routing protocols, specific IP addresses, or even applications such as HTTP or SMTP.In filtering operations, the traffic to be filtered is assigned to the Discard VLAN. Since the Discard VLAN is not configured to exit the switch, the traffic will be discarded.
The figure above shows a common configuration in which a routed backbone is using both RIP and OSPF for its routing protocols. The network administrator does not want the multicast OSPF and broadcast RIP frames propagated to the end stations. The network is designed so that only end users are attached to the Matrix devices.
To implement filtering in this scenario, a Layer 3 rule and a Layer 4 rule will be created.
Rule 1 (Layer 3) - Any frame received with an IP Protocol Type of 89 (OSPF) will be classified into the Discard VLAN.
Rule 2 (Layer 4) - Any frame received with a Bilateral UDP port number of 520 (RIP) will be classified into the Discard VLAN.
Based on this configuration, as long as the Discard VLAN is not configured to exit the switch, all RIP and OSPF frames will be filtered from the end users.
Traffic Security
Traffic Security uses the same concepts as Traffic Filtering. Imagine a scenario where network access is provided to a group of unknown users. There have been problems with these unknown users "hacking" into the router and altering the configuration. A simple classification rule can be put in place that will prevent these types of occurrences.
In the figure above, the network components include a router and a Matrix E7. In this configuration end-users connect to the ports of the Matrix E7.
Since the end-users would never need to communicate directly to the router using the router's IP address, a Layer 3 IP classification rule will be used.
Rule - Any frames received by the switch with a destination IP address of the router (129.168.1.2) will be classified to the Discard VLAN. The Discard VLAN is configured so that it does not exit the switch.
The end result is that any frames from a user trying to "hack" into the router will be discarded before ever reaching the router.
Traffic Prioritization
Classification rules can be used to specify that certain network applications receive the highest transmission priority. For example, a network administrator wants to assign priority to three network applications, SAP R/3, web traffic, and email, in that order.
To accomplish the prioritization goals in this example, there are two main steps required: creating the classification rules, and configuring the Priority-to-Traffic Queue mapping for the switch.
First, create one Layer 3 and two Layer 4 classification rules.
Rule 1, Layer 3 (SAP R/3) - All frames to or from the IP address of the SAP R/3 server will be tagged with a priority indicator of 7 (highest).
Rule 2, Layer 4 (Web) - All frames with a UDP port number of 80 (HTTP) will be tagged with a priority indicator of 5.
Rule 3, Layer 4 (email) - All frames with a UDP port number of 25 (SMTP) will be tagged with a priority indicator of 3.
Note: An IP address classification was selected for Rule 1 because it has been observed that SAP R/3 dynamically negotiates the TCP/UDP port used, so the port number selections vary from session to session. If this was not the case, a Layer 4 UDP classification could be used.
Then, configure the priority-to-traffic queue mappings. Based on the default priority-to-traffic queue mapping (shown in the following table), the priorities assigned above will work out so that each frame classification type will be mapped to the desired traffic queue. This means that no user configuration of the priority-traffic queue mapping would be required. However, this mapping can be configured by the network administrator using local management or NetSight Atlas Console.
This table shows the default Matrix E7 priority-to-queue mappings.
Priority Value | Traffic Queue |
---|---|
7 | 3 |
6 | 3 |
5 | 2 |
4 | 2 |
3 | 1 |
2 | 0 |
1 | 0 |
0 | 1 |
Priority 7 = Highest
Traffic Queue 3 is serviced first
With the classification rules described in Traffic Prioritization, the network traffic would be prioritized as shown in the table below.
Application | Classification Type | Desired Priority | Priority Value | SS6000 Traffic Queue |
---|---|---|---|---|
SAP R/3 | Bilateral IP | High | 7 | 3 |
Web | UDP Port Number | Medium | 5 | 2 |
UDP Port Number | Low | 3 | 1 |
Classification Precedence Rules
When there is a role with multiple classification rules assigned to a port, the device must determine which rule takes precedence. The order of precedence is predefined in the device, and is not configurable. Network administrators should have a comprehensive understanding of classification precedence, as it can significantly impact the operation of traffic classification rules. The Device Support Tab (Role) provides rule precedence information for each role.The device determines the order of precedence based on the classification types. The Precedence Table details the order of precedence with 1 being the highest precedence, and 15 being the lowest.
Classification | Precedence |
---|---|
Layer 2 | |
Source MAC Address | 1 |
Destination MAC Address | 2 |
Ethertype Field / DSAP/SSAP Fields | 13 |
Layer 3 | |
IP TOS / IPX COS | 11 |
IP Protocol Type / IPX Packet Type | 12 |
Source IP Address Exact Match / Source IPX Network Number | 3 |
Source IP Address Best Match / Destination IPX Network Number | 4 |
Destination IP Address Exact Match | 5 |
Destination IP Address Best Match | 6 |
IP Fragment | 7 |
Layer 3/4 | |
IPX Socket Source / UDP Source Port / TCP Source Port | 8 |
IPX Socket Destination / UDP Destination Port / TCP Destination Port | 9 |
ICMP | 10 |
VLAN ID | 14 |
Priority | 15 |
Exact Match indicates a match of an explicitly defined address.
Best Match indicates a match of an entire subnet, or range within a subnet.
Note: The precedence of a rule based on a bilateral address match is determined frame by frame depending on whether the rule matches the destination or source address in the frame. A bilateral address rule which matches the source address has higher precedence than a rule which matches a destination address (or any other lower precedence rule).
Precedence Scenarios
The following scenarios illustrate the classification precedence rules.Scenario 1
A network administrator has defined two classification rules:Rule 1- All frames with a UDP port number of 55(ISI Graphic Language) are assigned to the Red VLAN.
Rule 2- All frames sourced from the 132.181.28.x subnet are assigned to the Blue VLAN.
If a frame is received with a source address of 132.181.28.99 and a UDP port number of 55, the frame will be assigned to the Blue VLAN because as shown in the Precedence Table, a Layer 3 IP Address rule takes precedence over a Layer 4 rule.
Scenario 2
A network administrator defines two classification rules:Rule 1- All frames with an IP TOS value of AA are assigned a priority of 7.
Rule 2- All frames with a TCP port number of 80 (HTTP) are assigned a priority of 3.
If a frame is received with a TOS value of AA and a TCP port number of 80, the frame will be assigned a priority of 3, because as shown in the Precedence Table, TCP port number classifications take precedence over IP TOS classifications.
Related Information
For information on related tasks: For information on related windows:![Top Top](http://www.enterasys.com/support/manuals/Pol_Mgr1_6_0-web/docs/images/p_up.gif)