TN3270E Performance Analysis using ITF

Overview

An analysis of two TN3270E (RFC1647) clients was conducted using the InterNetworking Test Facility MS3270 tests. These tests are designed to evaluate TN3270 and TN3270E single and multi session performance under a variety of conditions. Note: The same type of analysis can be performed for SNA products using the SNA Development Test Facility (SDTF).

Test Environment

The tests were run using a 4 megabit Token Ring attachment to an ITF TN3270 server. Two TN3270 emulators were evaluated using 1 session and 4 sessions with a common TCP/IP stack. The emulators will be referred to as 'x' and 'y' in this report to protect the names of the innocent.

Categories of Test Performed

The following categories of test were run:

Outbound Only
Outbound/Inbound with Read Modified and Read Buffer
Inbound Only

The tests contained the following types of transactions:

sequential - transaction was completed to each session before accessing the next session
overlapped - transaction output and input to all sessions was asynchronous
no response - transaction output required no response

MS3270 Test Results

Outbound Only

Outbound/Inbound Read Modified

Outbound/Inbound Read Buffer

MS3270 Test Results Analysis

The ITF MS3270 tests were used to evaluate the performance of 1 session and 4 sessions for each emulator. Emulator 'x' was then compared with emulator 'y'.

Emulator 'x'

Outbound Only

Sequential output to 4 sessions was slightly slower than to 1session.
Overlapped performance was noticeably faster than no response and 4 sessions performed better than 1.
Overlapped performance was better than sequential or no response.

Outbound/Inbound Read Modified

Sequential output to 4 sessions was identical to 1 session.
Overlapped performance was faster than no response and 4 sessions performed better than 1.
Inbound 4 session performance was much better than 1 session.
Overlapped performance was better than sequential or no response.

Outbound/Inbound Read Buffer

Sequential output to 4 sessions was identical to 1 session.
Overlapped performance was faster than no response and 4 sessions performed better than 1.
Overlapped performance was better than sequential or no response.

Emulator 'y'

Outbound Only

Sequential output to 4 sessions was slightly slower than to 1session.
Overlapped performance for 1 session and 4 sessions was the same.
No response performance for 4 sessions was slightly better than 1.
Performance of sequential and overlapped was nearly identical and no response slower than either.

Outbound/Inbound Read Modified

Sequential output to 4 sessions was slightly slower than to 1session.
Overlapped performance for 1 session and 4 sessions was nearly the same.
No response performance for 4 sessions was worse than 1.
Performance of sequential and overlapped was nearly identical and no response slower than either.

Outbound/Inbound Read Buffer

Sequential output to 4 sessions was identical to 1 session.
Overlapped performance for 4 sessions was better than 1 session.
No response performance for 4 sessions was worse than 1.
Performance of sequential and overlapped was nearly identical and no response slower than either.

Emulator 'x' vs 'y'

Outbound Only

'x' performs better than 'y'.

Outbound/Inbound Read Modified

'y' is slower than 'x' for outbound/inbound.
'y' inbound performance is better or equal to that of 'x'.

Outbound/Inbound Read Buffer

'y' is slower than 'x' for all categories. The data gathered during testing is given below. The following notation is used:
TestID - is the test number
UBPS - is the actual bits/second of user data
BPS - is the actual bits/sedond of TCP/IP data and user data
PPS - is the number of packets/second
TPS - is the number of user transactions/second
RspTm - is the response time for each transaction

TN32Ex 1 session

TestID UBPS BPS PPS TPS RspTm

1 16,502 17,983 26.5 13.2 0.08

2 16,526 18,009 26.5 13.2 0.08

3 13,483 14,094 10.9 10.9 0.09

4 10,935 12,684 31.2 7.8 0.13

5 10,825 12,553 30.9 7.7 0.13

6 12,555 14,069 27.0 9.0 0.11

7 130,864 132,614 31.3 7.8 0.13

8 130,579 132,325 31.2 7.8 0.13

9 152,405 153,935 27.3 9.1 0.11

10 2,358 4,393 36.3 18.2 0.06

11 2,361 4,398 36.4 18.2 0.06

12 283,173 285,205 36.3 18.1 0.06

13 283,150 285,182 36.3 18.1 0.06

14 112,981 113,755 13.8 6.9 0.15

15 112,113 112,881 13.7 6.9 0.15

16 85,824 86,120 5.3 5.3 0.19

TN32Ex 4 sessions

TestID UBPS BPS PPS TPS RspTm

1 15,860 17,285 25.5 12.7 0.08

2 25,266 27,536 40.5 20.3 0.05

3 23,447 24,514 19.0 19.0 0.05

4 10,957 12,711 31.3 7.8 0.13

5 20,141 23,356 57.4 14.4 0.07

6 23,611 26,457 50.8 16.9 0.06

7 130,340 132,083 31.1 7.8 0.13

8 176,767 179,133 42.3 10.6 0.10

9 95,922 96,884 17.2 5.7 0.18

10 2,459 4,500 36.4 18.2 0.06

11 9,739 18,155 150.3 75.1 0.01

12 283,213 285,244 36.3 18.1 0.06

13 512,959 516,640 65.7 32.9 0.03

14 86,139 86,729 10.5 5.3 0.19

15 163,037 164,155 20.0 10.0 0.10

16 92,995 93,321 5.8 5.8 0.17

TN32Ey 1 session

TestID UBPS BPS PPS TPS RspTm

1 13,334 14,531 21.4 10.7 0.09

2 13,272 14,463 21.3 10.6 0.09

3 11,069 11,571 9.0 9.0 0.11

4 13,166 15,269 37.5 9.4 0.11

5 13,079 15,167 37.3 9.3 0.11

6 13,578 15,214 29.2 9.7 0.10

7 61,777 62,603 14.8 3.7 0.27

8 61,941 62,769 14.8 3.7 0.27

9 61,635 62,254 11.0 3.7 0.27

10 9,028 16,896 140.5 70.3 0.01

11 9,040 16,919 140.7 70.3 0.01

12 95,535 96,221 12.2 6.1 0.16

13 95,190 95,873 12.2 6.1 0.16

14 108,653 109,398 13.3 6.7 0.15

15 106,403 107,133 13.0 6.5 0.15

16 60,485 60,694 3.7 3.7 0.27

TN32Ey 4 sessions

TestID UBPS BPS PPS TPS RspTm

1 12,713 13,854 20.4 10.2 0.10

2 13,166 14,352 21.2 10.6 0.09

3 12,693 13,273 10.4 10.3 0.10

4 10,781 12,507 30.8 7.7 0.13

5 13,016 15,088 37.0 9.3 0.11

6 9,048 10,141 19.5 6.5 0.15

7 61,267 62,087 14.6 3.7 0.27

8 94,740 96,011 22.7 5.7 0.18

9 40,029 40,431 7.2 2.4 0.42

10 7,886 14,673 121.2 60.6 0.02

11 9,488 17,675 146.2 73.1 0.01

12 95,017 95,698 12.2 6.1 0.16

13 219,554 221,129 28.1 14.1 0.07

14 93,613 94,255 11.5 5.7 0.18

15 108,060 108,801 13.2 6.6 0.15

16 101,135 101,489 6.3 6.3 0.16

TestID	UBPS	BPS	PPS	TPS	RspTm
1	16,502	17,983	26.5	13.2	0.08
2	16,526	18,009	26.5	13.2	0.08
3	13,483	14,094	10.9	10.9	0.09
4	10,935	12,684	31.2	7.8	0.13
5	10,825	12,553	30.9	7.7	0.13
6	12,555	14,069	27.0	9.0	0.11
7	130,864	132,614	31.3	7.8	0.13
8	130,579	132,325	31.2	7.8	0.13
9	152,405	153,935	27.3	9.1	0.11
10	2,358	4,393	36.3	18.2	0.06
11	2,361	4,398	36.4	18.2	0.06
12	283,173	285,205	36.3	18.1	0.06
13	283,150	285,182	36.3	18.1	0.06
14	112,981	113,755	13.8	6.9	0.15
15	112,113	112,881	13.7	6.9	0.15
16	85,824	86,120	5.3	5.3	0.19