Traffic Squeezer - Performance Benchmarks

Traffic Squeezer - Performance Benchmarks

Here is the interesting real Traffic Squeezer Performance Benchmarks. Instead of static number attached the real screen-shots for the purpose. Also one can see the trend/pattern the way TS is performing with each test scenarios.

* To make things simple and to make more understandable, the performance of Traffic Squeezer is tested here mostly with one side Server (traffic source) and other side with a Client Application (traffic sink). The graph with more shaded area represents compression percentages. So obvious flow towards Server to Client is more optimized and the flow from Client to Server is less optimized. Or in other words, since client getting/fetching data from server. But this is not the case for certain benchmark tests where there is no client <> server scenario, for example VOIP/AV conference and so on.

HTTP Optimization Statistics
This is a part of Application Specific Optimization, which is HTTP Optimization in this case. Below which it shows the percentage of the optimization specific to HTTP Packets. Please note this is just HTTP Specific Optimization alone. The overall Optimization is always a summation of all optimization techniques put together. Or in other words, if you enable HTTP Optimization and Compression, then overall optimization is the summation of both HTTP Optimization (which the way the graph shows below, and Compression graph.

Compression - transition from HTTP to VNC (Remote-Desktop) Protocols
Here is a interesting migration where the Compression technique degrades, when a the traffic pattern is changed, such as from HTTP (which can be highly optimized) to VNC session. VNC session (being a Remote Desktop) protocol, already sends traffic which is highly compressed (VNC mostly uses ZRLE which is already compressed), hence the Compression Ratios in Traffic Squeezer suddenly falls.

Compression - MySQL Client to MySQL Remote Server Session
A MySQL Client and MySQL Remote Server session is established. To test scenario various standard SQL DML and DDL based commands are executed. Also query is executed upon huge volumes of data using SELECT with large number of rows as query result criteria.

Compression - NetMeeting (VOIP/Audio + White-board Sharing) Session
This is a simple end to end NetMeeting based test scenario. It has been tested with Audio Conferencing (so VOIP) as well White-board sharing tool. This test has been done on Compression alone, since VOIP packets are small UDP packets, one can achieve higher optimization only with Coalescing feature too enabled. With compression alone, since the packets are already small there is not much room for compression in this case. Also please note unlike other test results, this scenario involves no data-source (server) and data-sink (client) based scenario, instead both end-to-end machines receives as well sends interactive data, being in a interactive conferencing mode.