Tài liệu KRONE - White paper - Zero Bit Error Rate doc

ZERO BIT ERROR RATE On Data Throughput in Structured Cabling Systems Released by KRONE Australia Asia Pacific Test Laboratory (Certified by Underwriters Laboratories) February, 2001 WHITE PAPER More and more network managers are demanding that their networks operate to their maximum potential, rather than merely operate as a standards- compliant solution. Downtime or a slowdown in network performance can be crippling to an organisation in today’s global e-commerce environment. The network cabling infrastructure must be able to support an enterprise’s increasing rates of data transmission now and in the future. It’s imperative that mission-critical information which is transmitted in packages of bits is delivered immediately, without costly delays. Bit Error Rate (BER) testing is one way of measuring a network’s performance - of seeing how quickly and efficiently information is delivered from sender to receiver. Zero Bit Error Rate (ZBER) has been the goal that now is a practical reality, thanks to KRONE’s TrueNet solution. However, controversy reigns in the cabling industry today disputing the ability to actually measure an absolute zero. Amid claims that ZBER is no more than a new buzz-acronym and, therefore, cannot be achieved, this White Paper outlines KRONE’s methodology of testing for, achieving and, indeed, guaranteeing Zero Bit Error Rate. WHAT THE STANDARDS SAY The IEEE 802.3 Standard defines the lowest acceptable value for Bit Error Rate (BER) for 1000 Base-T as being 10 - 10 . This BER limit is set for levels of line encoding using pulse amplitude modulation (PAM-5). This is the encoding scheme for Gigabit Ethernet. WHAT IS BER? A bit error occurs when the transmitted signal is corrupted by an event which changes the voltage level of the signal. Events such as noise, impulses due to switching surge or lightning, transmission equipment fault plus others can all cause a bit error (eg. it is the reception of 0 when 1 was transmitted or vice versa). The Bit Error Rate (BER) is a measure of how often these errors occur. Statistical Method Some measurement techniques are based on statistical analysis which predict the number of bit errors with a certain degree of accuracy using common statistical distribution methods. For example: Pb(e) is bit error probability which equals BER. P(e) is the probability that a waveform is detected incorrectly. M is the number of discrete signal levels (eg 3 for PAM-3, 5 for PAM-5). Practical Method Although KRONE uses both statistical and practical methods, we put more emphasis on the measurement of real errors as a way of understanding the nature of an error. Any modern LAN hardware and network analysis software can perform this real physical measurement. Practical measurement techniques use the cyclic redundancy check (CRC) to determine actual bit error failures in any period of time. For example, there may be two CRC failures after running a system using PAM-5 encoding for, say, 100 hours. CRC uses a frame check sequence, developed by the transmitter, that is checked by the receiver to see if it is valid or not. If it is not valid, because at least one bit contains an error, the receiver rejects the ZERO BIT ERROR RATE Cabling infrastructure providers are following KRONE’s lead regarding the importance of having the lowest bit error rate possible. This white paper confirms that zero bit error rate is a practical reality that Network managers can rely on. P (e)_ Log 2 M < Pb (e) < P (e) whole packet of data that may contain up to 1500 bytes x 8 bits = 12,000 bits for an Ethernet Frame. This means that one bit error can cause up to 12,000 bits to be re-transmitted. WHAT IS A ZERO BIT ERROR RATE? IEEE 802.3 defines the occurrence of errors at 10- 10 as being its worst case. At this condition any bit errors which may exist do not degrade the performance of the network because all network hardware and software is built to meet this condition. Therefore, any noise present at this condition will be insufficient to alter the bit value at the receiver and cause an error. KRONE has selected a level 100 times lower than the worst case in the Standard and has called this BER condition of 10- 12 as ZERO BIT ERROR RATE. Zero Bit Error Rate (ZBER) means less than one error in a million-million bits. WHAT KRONE DOES In-House Testing We consider field network testing to be the most accurate way to correlate theoretical base to real world performance. To do this, we measure the product performance in the laboratory under severe noise conditions (simulating worst case real life environment) and then we test the installed cabling systems at the customer’s site to verify the laboratory performance in no-load and live network conditions (eg the network is fully functional). When testing, KRONE uses the latest test technologies and the most sophisticated digital measurement equipment available on the market such as Netcom Systems Smartbits SMB-2000, Vigilant Big Tangerine, and Fluke DSP 4000. These tools are recognised as the best by all international test laboratories such as The Tolly Group, BiCSi, Bell Laboratories, etc. In the example below, KRONE used Smartbits SMB- 2000 Signal Generators generating up to 19 Gigabits of continuous traffic in both directions at once. A Pentium 200MHz laptop computer was used to monitor the traffic for four days while the site was influenced by neighbouring external noise sources. There were zero delivery errors. KRONE went even further, analysing network latency using advanced Netcom SmartFlow v. 1.14 network analysis software generating different groups of data streams. The results showed that during the tests no frames were lost on KRONE’s TrueNet ™ cabling system. Field Testing For field testing of commercial and industrial sites, KRONE uses an advanced 24-port Vigilant Network Monitor to evaluate the health and integrity of a customer’s structured cabling system on a continuous basis. From this testing, measurements of impedance versus distance and frequency are made, finding the impedance mismatches and discontinuities in a network channel that cause network traffic to slow down. We found that impedance mismatch was the main factor that caused performance loss in Test 1 Test 2 Test 3 Gigabit Card No 01 GX-1420B 01 GX-1420B 01 GX-1420B Tx Frames 237,644,979 13,115,406 11,393,017 Rx Frames 233,998,500 12,648,131 11,631,058 Rx Bits 14,975,904,000 19,199,862,858 17,655,946,044 Rx Triggers 0 0 0 CRC Errors 0 0 0 Over Size 0 0 0 Frag/Under Size 0 0 0 BER 000 TrueNet C6T KRONE Platinum Solution a structured cabling system. TIA/EAI-568A standard specifies the impedance variation to be +/-15 Ohms but KRONE’s TrueNet ™ Cabling System goes even further, restricting it to ± 3 Ohms. The test equipment captures noise, interference and cross-talk events in the network and analyses the effect they have on performance. From these measure-ments we have been able to verify that correctly installed TrueNet ™ Cabling Systems have impedance which is within ± 3 Ohms throughout the cabling channel. The graph below shows test results from a TrueNet cable run in a customer’s structured cabling system. Most impedance mismatch problems happen in the cross-connect patch cord in the closet, the consolidation point, and the connectors at the telecommunications outlet. The first few metres on the graph relate to the temporary use of the injecting patch cord - which is the connection between the tester and the cable run. The rest of the graph shows that the cable run and all components are within ± 3 Ohms resulting in negligible impedance mismatches that will not cause bit errors and thereby data retransmissions. THE RESULT Both in-house testing and on-site testing confirm KRONE’s technical approach to proving the feasibility of achieving Zero Bit Error Rates is correct. More importantly, KRONE customers confirm that Zero Bit Error Rates are achievable on their installed sites. Because KRONE components are tuned to remove impedance mismatches, the TrueNet ™ Cabling System does perform with a bit error rate lower than 10- 12 . This well exceeds the standards requirements of 10- 10 . Therefore, KRONE can offer a 5-year “ZBER” throughput warranty to sites that install a KRONE TrueNet ™ structured cabling system. For further information please contact: KRONE (Australia) Holdings Pty Limited PO Box 335, Wyong NSW 2259 2 Hereford Street, Berkeley Vale NSW 2261 Tel: 02 4388 4422 Fax: 02 4388 4499 Email: kronehlp@krone.com.au Web: www.krone.com.au 5549 10/01 KRONE (N.Z.) Technique Limited PO Box 38-177 Wellington Mail Centre Wellington, New Zealand Tel: 0800 657 663 Fax: 0800 355 100 Email: sales@krone.co.nz Berkeley Vale, Australia February, 2001 By Boris Torban Peter Meijer Telecommunications Engineer Technical Training Manager KRONE (Australia) Holdings Pty Limited KRONE (Australia) Holdings Pty Limited Injecting patchcord Cable run . BER condition of 1 0- 12 as ZERO BIT ERROR RATE. Zero Bit Error Rate (ZBER) means less than one error in a million-million bits. WHAT KRONE DOES In-House Testing We. 8 bits = 12,000 bits for an Ethernet Frame. This means that one bit error can cause up to 12,000 bits to be re-transmitted. WHAT IS A ZERO BIT ERROR RATE? IEEE