Downloads |
Document ID: 21974
Contents
Introduction
Before You Begin
Conventions
Prerequisites
Components Used
Determine Savings
Proxy Caching
Transparent Caching
Using a Router to Determine Bandwidth Savings
Explanation of the Show Statistics Commands
Related Information
Introduction
This document shows how cache savings can be calculated using the tools and commands available on Cache Engines, Content Engines, and Routers.
Before You Begin
Conventions
For more information on document conventions, see the Cisco Technical Tips Conventions.
Prerequisites
Readers of this document should be knowledgeable of the following:
-
Transparent and Proxy Caching
-
Web Cache Communication Protocol (WCCP)
-
Cisco IOS©
-
Hypertext Transfer Protocol (HTTP)
Components Used
The information in this document is based on the software and hardware versions below.
-
Cisco IOS 12.1 through to 12.2.7
-
Caching - Content Software 2.5.1 and ACNS 4.X
-
All Route Platforms
-
All Caching - Content Platform CE507, CE560, CE590, CE7320, CE505, CE550, CE570
Determine Savings
There are three areas of discussion when looking at the savings available for caching; transparent caching, proxy caching and using the routers interface counters.
Proxy Caching
In a proxy caching setup where the clients browses are set to point directly to the cache engine on a predetermined port the calculation of savings is straight forward.
There are different commands that are used to determine exactly how the cache is performing.
show statistics http savings
show statistics ftp
show statistics wmt. savings (Only available in ACNS 4.X and above)
show statistics mediacache real savings (Only available in ACNS 4.X and above)
Note: Although you can set the secure proxy (https) up in the proxy settings on a browser we only tunnel this session and can not cache it because it is encrypted.
Transparent Caching
In a transparent caching setup, where no settings are required in the clients browser and traffic is intercepted at a router and sent to the cache using the Web Cache Communications Protocol(WCCP), the calculation of saving is just as straight forward.
These are the commands needed to determine the savings.
show statistics http savings
show statistics wmt savings (Only available in ACNS 4.X and above)
show statistics mediacache real savings (Only available in ACNS 4.X and above)
Note: In transparent mode we can not cache ftp or https sessions. These sessions are not sent to the cache and are passed directly to the origin server.
Using a Router to Determine Bandwidth Savings
When using a router to determine the saving there are a number of things to consider to get an accurate measure of the savings in raw bandwidth. The most important thing to do is get a baseline. In order to do this you need to monitor the link utilization of the upstream link from your router that is connected to the internet. In a reverse proxy caching (RPC) setup you need to monitor the load on your servers. There are many tools available to monitor link utilization via SNMP. On freely available tool is MRTG, this application is not supported or supplied by Cisco. It can be found at Multi Router Traffic Grapher 
.
On a Cisco router you monitor the following counters on the show interface command on the link to the internet, bytes in and bytes out. In order to get a base line you need to understand the makeup of the traffic heading out to the internet. In transparent mode only http, wmt(1) and real(1) requests are sent to the CE. While in proxy mode ftp(2) and https(3) can be sent to the cache as well as http, wmt(4) and real(4). A traffic analyzer is an invaluable tool in this situation. Ports used for the protocols mentioned are as follows. These are the default port and can be changed.
http tcp 80 https tcp 443 ftp tcp 20 and 21 wmt tcp 1755 or udp 1755 or http or multicast real rtsp 554
(1) Only if configured as a WCCP service.
(2) This only applies when the ftp session is down from within a browser or application that has the ftp proxy set to the cache.
(3) Although HTTPS (Secure) sessions are sent to the cache, they can not be cache and are only tunneled through.
(4) The proxy setting needs to be set within the Windows Media Player or the Real Player. Please consult the user's guide for information on how to set these.
The methodology for calculation of savings using interface counters is as follows:
-
Without caching turned on clear the interface counters for the connection to the internet (clear counters is the command for Cisco IOS).
-
Wait for a 2 average days and note down the highlighted counter values below. Note the time that the counters were running for.
-
Turn your caching on, whether it be proxy or transparent caching,. You need to wait for the cache to populate. This should be about 3 days depending on load. Savings should been seen after a few hours, but three days will give you a good base line.
-
Clear the interface counters on the connection to the internet again.
-
With the caching turned on wait the same amount of time as with caching turned off.
-
Note the interface byte counters again.
-
The difference is your bandwidth savings. This will be a different percentage compared to the sh stat http saving due to the router interface seeing all traffic while the cache only sees the protocols that are going to be cached.
Note: The below stats are from a router that is doing only http and hence the good figures. On a production router this would be far less.
Note: Show interface without caching.
Serial0/0 is up, line protocol is up Hardware is PowerQUICC Serial Internet address is 10.64.21.10/30 MTU 1500 bytes, BW 128 Kbit, DLY 20000 usec, reliability 255/255, txload 59/255, rxload 194/255 Encapsulation HDLC, loopback not set Keepalive set (10 sec) Last input 00:00:00, output 00:00:00, output hang never Last clearing of "show interface" counters 00:06:52 Input queue: 4/75/0/0 (size/max/drops/flushes); Total output drops: 6 Queueing strategy: weighted fair Output queue: 0/1000/64/6 (size/max total/threshold/drops) Conversations 0/32/32 (active/max active/max total) Reserved Conversations 0/0 (allocated/max allocated) Available Bandwidth 96 kilobits/sec 5 minute input rate 177000 bits/sec, 47 packets/sec 5 minute output rate 30000 bits/sec, 44 packets/sec 14218 packets input, 8743319 bytes, 0 no buffer Received 42 broadcasts, 0 runts, 0 giants, 0 throttles 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored, 0 abort 13019 packets output, 1113797 bytes, 0 underruns 0 output errors, 0 collisions, 0 interface resets 0 output buffer failures, 0 output buffers swapped out 0 carrier transitions DCD=up DSR=up DTR=up RTS=up CTS=up
Note: Show interface with caching turned on.
Serial0/0 is up, line protocol is up Hardware is PowerQUICC Serial Internet address is 10.64.21.10/30 MTU 1500 bytes, BW 128 Kbit, DLY 20000 usec, reliability 255/255, txload 1/255, rxload 1/255 Encapsulation HDLC, loopback not set Keepalive set (10 sec) Last input 00:00:06, output 00:00:07, output hang never Last clearing of "show interface" counters 00:17:33 Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0 Queueing strategy: weighted fair Output queue: 0/1000/64/0 (size/max total/threshold/drops) Conversations 0/11/32 (active/max active/max total) Reserved Conversations 0/0 (allocated/max allocated) Available Bandwidth 96 kilobits/sec 5 minute input rate 0 bits/sec, 0 packets/sec 5 minute output rate 0 bits/sec, 0 packets/sec 7503 packets input, 5408948 bytes, 0 no buffer Received 105 broadcasts, 0 runts, 0 giants, 0 throttles 0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored, 0 abort 5723 packets output, 497401 bytes, 0 underruns 0 output errors, 0 collisions, 0 interface resets 0 output buffer failures, 0 output buffers swapped out 0 carrier transitions DCD=up DSR=up DTR=up RTS=up CTS=up
Explanation of the Show Statistics Commands
Show statistics http savings
cache-1#show statistics http savings Statistics - Savings Requests Bytes ----------------------------------------------------------- Total: 663699 5369304364 Hits: 261820 583380690 Miss: 401782 4612002188 Savings: 39.4 % 10.9 %
The percentages are the figures that give you the best indication. As you can see from the above output, we have requests savings of 39.4% and byte savings of 10%. This is within the expected range of savings.
Show statistics ftp
cache-1#show statistics ftp FTP Statistics -------------- FTP requests Received = 6 FTP Hits Requests Percentage Number of hits = 3 50.0 % Bytes = 54171214 50.0 % FTP Misses Requests Percentage Number of misses = 3 50.0 % Bytes = 54171214 50.0 % Requests sent to Outgoing Proxy = 0 Requests sent to origin ftp server = 3 FTP error count = 0
Show statistics wmt savings
cache-1#show statistics wmt savings Unicast Savings Statistics ========================== Total bytes saved: 15271392 -------------------------- Total % of Total Bytes Saved -------------------------------------------- By Pre-positioned content: 0 0.00% By Live-splitting: 11770875 77.08% By Cache-hit: 3500517 22.92% Total % of Total Live Outgoing Bytes -------------------------------------------- Live Splitting -------------- Incoming bytes: 26367513 69.14% Outgoing bytes: 38138388 100.00% Bytes saved: 11770875 30.86% Total % of Bytes Cache Total -------------------------------------------- Caching ------- Bytes cache-miss: 10761747 75.46% Bytes cache-hit: 3500517 24.54% Bytes cache-total: 14262264 100.00% Bytes cache-bypassed: 0 Total % of Req Cache Total -------------------------------------------- Cacheable requests ------------------ Req cache-miss: 1 16.67% Req cache-hit: 1 16.67% Req cache-partial-hit: 4 66.67% Req cache-total: 6 100.00% Req cache-bypassed: 0 Objects not cached ------------------ Cache bypassed: 0 Exceed max-size: 0
Show statistics mediacache real savings
ce-7320-cache#show statistics mediacache real savings Media Cache Statistics - Savings Requests Bytes ----------------------------------------------------------- Total: 0 0 Hits: 0 0 Miss: 0 0 Savings: 0.0 % 0.0 %
Related Information
- WCCP V1
- WCCP V2
- ACNS Documentation
- Caching Documentation
- WCCP Commands
- Technical Support - Cisco Systems
| Updated: Jan 30, 2006 | Document ID: 21974 |