RFC2544/NE Testing
RFC2544 Test Notes
Test Layers
RFC2544 is designed for Layer 2 and Layer 3 devices. As such, each test frame must
have a valid MAC header, preamble, and interpacket gap. For testing Layer 3
devices, such as routers, a valid IP header is also required.
Though VLAN support is not mentioned in RFC 2544, VLAN-based services should
include the appropriate VLAN tags. Unframed testing, where the payload data is not
encapsulated into a valid Ethernet frame, is not compatible with RFC2544 device
testing.
XTT 5000 uses a frame payload that consists of a sequence number, a time stamp,
and a test pattern specified by the user. The sequence number and time stamp are
used to accurately measure lost frames and latency, respectively.
Frame Sizes
The standard frame sizes for Ethernet testing are 64, 128, 256, 512, 1024, 1280,
and 1518 bytes. With systems that support jumbo frames, such as 4096- or 9000-
byte frames, these frame sizes should be tested as well. The XTT 5000 defaults to
the frame size defined in RFC 2544, but allows you to set the frame size to any valid
value.
RFC2544 Test Types
Throughput
The throughput test determines the maximum frame rate without lost frames the
DUT can manage. The test begins at 100% frame rate by sending a predetermined
number of frames, or, more commonly, sending the frames for a predetermined
length of time. If any frames are lost, the test is repeated at a lower frame rate. This
process continues until the maximum throughput is determined.
Sunrise Telecom uses a binary search algorithm for determining the throughput. The
standard test method reduces the throughput by a set increment, such as 10%. This
is not the most efficient algorithm available especially for determining the throughput
with a better resolution, such as 1%. The binary search changes the throughput
value by ever decreasing increments: 50%, 25%, etc. The throughput is increased or
decreased depending on the results of the previous test. The algorithm continues to
run until the throughput is determined to within the specified resolution, typically 1%
to 10%.
Latency
The standard latency test is to run test traffic at the predetermined throughput rate
or two minutes, and measure the latency of a single tagged frame sent at least one
minute into test. The reported latency is the average of twenty such tests.
Strict adherence to the standard would require 280 minutes, over four hours, to
complete for all frame sizes. The XTT 5000 provides the option to instead perform a
'Quick Latency” test that eliminates the need to run a separate and time consuming
latency test. During the throughput test, the latency of the test frames is measured
87
Summary of Contents for XTT 5000
Page 1: ...XTT 5000 User s Manual...
Page 2: ......
Page 6: ......
Page 36: ......
Page 55: ...Throughput Testing L2 BERT Diagram Home Layer 2 BERT Where do you want to go next Home 49...
Page 58: ...XTT 5000 User s Manual Throughput Stream Results Throughput Test Setup Working Desktop Home 52...
Page 61: ...Throughput Testing Home 55...
Page 67: ...Throughput Testing Throughput Aggregate Results Throughput Stream Results Home 61...
Page 81: ...Throughput Testing Working Desktop Home 75...
Page 83: ...Throughput Testing Home 77...
Page 116: ......
Page 142: ......
Page 147: ...Save Features 141 RFC2544 NE Applications IP Test Setup Monitor Setup Loopback Test Setup Home...
Page 148: ......
Page 172: ......
