KVM Cable Length Best Practices Guide

Transcription

1 Infrastructure Management & Monitoring for Business-Critical Continuity TM KVM Cable Length Best Practices Guide What Customers Need to Know About Cable Length and Video Quality

2 Cable Length and Video Quality Table of Contents 2 Executive Summary 3 Overview of Cable Length and Video Quality 4 Video Quality Test Results 5 Conclusion Executive Summary Avocent has an ongoing goal to provide and ensure best-in-class solutions for IT management. As part of this goal, Avocent has historically chosen a maximum KVM cable length of approximately 33 feet (10 meters). This number was selected because KVM will work in that range, regardless of the conditions. In reality, there is actually no specific cable length at which video will just stop working. There are, instead, several factors that combine to affect the video quality. Cables longer than 33 feet (10 meters) may be used if the customer s environment is carefully made to fit as closely as possible to the ideal conditions set forth in this document. 2 Cable Length and Video Quality

3 Overview of Cable Length and Video Quality When cable longer than Avocent s recommended maximum length is used, video does not just suddenly stop working. Instead, it will begin slowly degrading. Video quality gradually degrades at the local port, and digitizing the signal just causes even more problems. The point at which the video actually becomes unusable is subjective. Some users may be more sensitive to sluggish or jumpy mouse movements than others, for example. Avocent s definition of acceptable video quality is video that is considered visually identical to the source using the minimum network bandwith to transmit changes and producing zero packets generated on static screens. The key factor in the amount of network traffic is the amount of changes in the video. These changes are recognized by the Avocent Dambrackas Video Compression (DVC) algorithm, which compares every single pixel from one frame to the next. Every recognized pixel change is packaged and sent from the appliance to the remote user. As network activity increases, mouse latency increases, which means that mouse movement may become sluggish or jumpy. Video noise is another key contributor to the level of network traffic. Theoretically, if a screen does not change, no packets should be generated. In reality, video noise causes some pixels to appear to change, even though there was no actual change. The appliance recognizes video noise as a screen change and sends those packaged pixel changes, thus generating more network traffic. Factors Affecting Video Quality There are seven variables that combine to affect video quality and network traffic: quality of video source, quality of cable, twist ratio of the cable, resolution used, length of the cable, number of patch panels and interference. To achieve the best possible video quality, careful attention must be shown to these factors. Quality of video source. Lower quality video cards will usually result in a noisier video source, and thus more network traffic, even on a static screen. Although the overall quality of video cards has improved over time, there are still some poor quality video cards being used. In addition to using better quality video cards, video noise can be reduced by enabling the video noise compensation feature, an appliance setting in the DSR switch. This feature tells the digital subsystem to raise the threshold for what counts as a change in video, basically eliminating any very slight video changes. Keep in mind, however, that if these minor changes in the target video are not detected, remote users will not see those minor changes either. Quality of the cable. Poor quality cable can increase video noise in the signal path as well. Therefore, Avocent recommends using CAT-5e and CAT-6 UTP cabling. With all shielded cable, the maximum distance is halved. For this reason, UTP is the preferred cable for AMX. In addition, you must consider not only the quality of the cable itself, but also the quality of the end connections. If the end connectors are not properly terminated, you may experience lower video quality and higher network traffic. Twist ratio of the cable. Twist ratio is how many twists per foot are on each pair of wires in the cable. Removing the individual wires, straightening them out and laying them side by side will show that none of the wires are the same length. The longer the length of the cable, the greater the difference in the length of separate wires. This difference creates a different delay for each of the three colors in the video. If the difference in delay becomes great enough, the three colors can be out of phase and seen in the video as ghosting or smearing. Resolution used. The frequency of the carrier for the video signal is higher on a 1280 x 1024 source than it would be on a 1024 x 768 source. Higher frequencies are more susceptible to noise and degradation over cable length. By using lower resolution, video signals can travel further down the cable before becoming too noisy to use. Length of the cable. The longer the cable, the more the capacitance of the transmission line and the bigger the difference in length of the pairs. This contributes to the degrading of the signal and the greater variation in delay in the three colors, a direct result of the length of the cable and twist ratio. However, as was seen in Avocent s testing, even the best cable will degrade the video signals at very great lengths. Number of patch panels. Each patch panel will require the cable to be terminated with an RJ connector. This involves taking the signal off of the twisted pair and putting it onto the blade of the male RJ, through the female RJ panel, and back onto a twisted pair either on another 4-pair wire or possibly larger pair bundles in structured cabling. Each time the signal is moved from the cable through a patch panel back into a cable, additional noise and crosstalk are generated. With Virtual Media products, the data channel on the last two wires is a much higher frequency than the non-virtual Media product. This additional high frequency can contribute noise and cross talk on the video lines - especially through a lot of patch panels. Interference. Electrical interference signals can also affect signal quality. Customers need to be aware of the potential negative effect of placing KVM cabling near sources such as AC units, power generators and fluorescent lights. 3 Cable Length and Video Quality

4 Video Quality Test Results Recently, tests performed by Avocent have prompted a change in the maximum cable length number from 33 feet (10 meters) to 150 feet (45 meters). Avocent has determined that cables up to this length, in an environment with carefully maintained variables, will work without significant reductions in video quality or resolution or increases in network traffic. Each of those capabilities was tested at six different cable lengths and was given a rating from one to four. As the graphs show, most scores remained the same or very close to the same, even for cable lengths as long as 500 feet. The exceptions are video quality and Virtual Media quality. They decrease in performance at a cable length of 150 feet (45 meters). This is why Avocent has determined that a new number of 150 feet (45 meters) is an acceptable cable length. It is extremely important to remember, however, that these tests were performed in a controlled testing environment, as discussed in Factors Affecting Video Quality, beginning on page 2. Window drag, minimize time and maximize time at local and remote ports. As shown in Figure 1.2, out of window drag, minimize time and maximize time, the window drag quality at remote port was the only quality adversely affected by cable lengths over 150 feet (45 meters). Window drag is a large video change, creating a lot of extra network traffic, which, combined with mouse movements, creates a network bottleneck. As a result, the user may see a perceivable delay in screen updates when using cable over 150 feet (45 meters). Figure 1:2 : Window Quality Test Results Keyboard, video and mouse capabilities at local port or during a DSView management software session. As shown in Figure 1.1, there was little to no change in mouse or keyboard quality at any cable length. However, at cable lengths of over 150 feet (45 meters), video quality decreased. The video was usable, but there was some smearing and tearing of text, and smaller details became more difficult to view. Figure 1:1 : KVM Quality Test Results 4 Cable Length and Video Quality

5 Virtual Media quality at local port or during a DSView 3 software session. As shown in Figure 1.3, Avocent s tests found that at cable lengths greater than 150 feet (45 meters), the signal latency in the high-speed data channel used for Virtual Media was too high. As a result, the Virtual Media data channel cannot be guaranteed to work when cable lengths are greater than 150 feet (45 meters). At that point, the DSR switch reverts to normal KVM data channel speed, so that the keyboard, video and mouse will still work properly. Figure 1:3 : Virtual Media Quality Test Results Conclusion Determining the guaranteed operation of transporting a video signal across a particular length of cable is a very complex issue. Historically, a cable length of 33 feet (10 meters) resulted in what Avocent considered to be acceptable video with acceptable network traffic regardless of the environment. More recent test results show that acceptable video can be achieved by using up to 150 feet (45 meters) with carefully controlled variables. As the cable length continues to increase, the perceived quality of the remote video will gradually begin to degrade at some point or create more network traffic, resulting in increased mouse latency. Because of the multiple factors that can affect performance at various cable lengths and the subjectivity of video performance, KVM configuration at cable lengths longer than the recommended 150 feet (45 meters) is not guaranteed by Avocent and should be supported only after a customer conducts thorough tests and achieves positive results from their own particular environment. 5 Cable Length and Video Quality

6 About Emerson Network Power Emerson Network Power, a business of Emerson (NYSE:EMR), is the global leader in enabling Business-Critical Continuity from grid to chip for telecommunication networks, data centers, health care and industrial facilities. Emerson Network Power provides innovative solutions and expertise in areas including AC and DC power and precision cooling systems, embedded computing and power, integrated racks and enclosures, power switching and controls, monitoring and connectivity. All solutions are supported globally by local Emerson Network Power service technicians. Aperture and Avocent solutions from Emerson Network Power simplify data center infrastructure management by maximizing computing capacity and lowering costs while enabling the data center to operate at peak performance. For more information, visit or Emerson Network Power. The global leader in enabling Business-Critical Continuity TM. AC Power Connectivity DC Power Embedded Computing Embedded Power Infrastructure Management & Monitoring Outside Plant Power Switching & Controls Precision Cooling EmersonNetworkPower.com Racks & Integrated Cabinets Services Surge Protection Emerson, Business-Critical Continuity and Emerson Network Power are trademarks of Emerson Electric Co. or one of its affiliated companies XX Emerson Electric Co. E KVMCL-TB-EN 000