Patent Application
20070099493
This invention relates to patch panels for computer systems and more particularly to the manner in which patch panels are laid out.
A patch panel is generally a panel of ports that can connect various different pieces of equipment together. Patch panels can allow circuits, networks, servers, and other devices to be arranged and rearranged, connecting the various equipment via ports and cables. For example, patch panels can connect a local area network's (LAN's) servers and/or computers to each other and to the outside lines that enable the LAN to connect to the Internet or another wide area network (WAN).
Many servers/computers to be connected are provided as rack-mountable devices. For example, servers are commonly provided as 19-inch rack equipment that can be mounted on 19 inch-wide racks (see e.g., Electronic Industries Alliance 310-D, International Electrotechnical Commission 60297, and Deutsches Institut für Normung 41494 SC48D). Various different devices are designed to be placed in racks and are commonly referred to as rack-mounts, rack-mounted systems, rack mount chasses, sub-rack, shelf, and the like.
Rack heights are typically of dimensions from about 77 inches to about 84 inches, the widths are typically from about 19 inches to about 28 inches, and the depths range from about 24 inches to about 42 inches. However, other rack sizes are also possible.
The rack's mounting fixtures include at least two parallel metal strips, also referred to as rails, standing vertically. Generally, for the rack size described above, each strip is about 0.625 inches (15.875 mm) wide, separated by a gap of about 17.75 inches (450.85 mm), giving an overall rack width of about 19 inches (482.6 mm). The strips also generally have holes in them at regular intervals so that each hole is part of a horizontal pair with a center-to-center distance of about 18.3 inches (464.82 mm).
The holes in the strips are generally arranged vertically in repeating sets of three, with center-to-center separations of 0.5 inch (12.7 mm), 0.625 inch (15.875 mm), 0.625 inch (15.875 mm). The hole pattern thus repeats every 1.75 inches (44.45 mm).
Racks are also divided into regions, generally 1.75 inches in height, within which there are three complete hole pairs in a vertically symmetric pattern, the holes being centered 0.25 inch (6.35 mm), 0.875 inch (22.225 mm), and 1.5 inch (38.1 mm) from the top or bottom of the region. Such a region is commonly known as a “U”, for “unit”, and heights within racks are measured by this unit. Rack-mountable equipment is generally designed to occupy some integral number of U. For example, rack-mountable computers are most often 1U or 2U high.
Rack-mountable equipment (i.e., servers and the like) can be mounted simply by bolting its front panel to the rack or, with a square-holed rack, by clipping or some other variation of fastener. In some cases, a second pair of mounting strips located at the back of the equipment can be utilized to mount the equipment to the racks as well. Various different spacing between the front and back strips is used (e.g., 800 mm can be a typical spacing), and equipment is often designed to handle a range of rack depths.
Typically, patch panels are horizontally arranged at the top of the server racks and usually take up approximately 1U per 24 ports of connections. In addition, an extra 2U may be taken up by cable runs or spacers for the cables.
According to some aspects, the present invention provides patch panels that can be vertically oriented. By moving the patch panels to a vertical orientation, not only is space saved, but cabling is also simplified.
In certain embodiments, the present invention provides patch panels that include panels having a plurality of network ports for ingoing and outgoing lines of a communications system. The panels can have a vertical length that is greater than a horizontal width and can include a plurality of mounting holes for rack units oriented along the vertical length. The patch panels of the present invention are capable of being vertically mounted on a rack.
In certain embodiments, the present invention provides methods for connecting two or more devices mounted on a server rack. Methods of the present invention can include connecting one or more cords to a patch panel having a plurality of network ports and vertically mounting the patch panel to the server rack.
In certain embodiments, the present invention provides an apparatus that can include a panel having a plurality of network ports for communication lines of a computer system containing one or more servers, where the plurality of network ports can be grouped together in a set of blocks. In certain embodiments, the panel can include mounting fittings for vertically mounting the panel to a server rack and mounting holes for rack units capable of being oriented along a vertical length of the panel where the vertical length of the panel is greater than a horizontal width of the panel. In certain embodiments the patch panel may include a hinge to allow for access to the rear of the panel after mounting.
Each of the figures diagrammatically illustrates aspects of the invention. The illustrations provide examples of the invention as described herein. Combinations of the aspects of specific variations or combinations of the specific variations themselves are within the scope of this disclosure.
A vertical patch panel, according to certain embodiments of the present invention, includes a plurality of network ports in a vertically oriented panel. The patch panel can be mounted vertically on the side of a server rack such that the patch panel can service all of the servers in the server rack. Changing the layout of the patch panel from horizontal to vertical, as in the invention, provides several advantages. A vertical patch panel that spans the length of the server rack simplifies cabling. The access to ports is moved closer to the servers. Shorter cabling is required, providing an overall cost savings as well as aesthetic benefits. By moving the patch panel to a vertical plane, a minimum of two rack units in the rack are saved. That is, a horizontal patch panel takes up two or more rack units, whereas the vertical patch panel does not take up any rack units space in the rack server.
Environment for use of Patch Panel
The patch panel of the invention can be mounted to a rack for servers. A server can generally be a device (e.g., a computer) on a network that manages network resources, although the patch panel of the invention can be used for servers having different functionality as well. For example, a file server is a computer and storage device dedicated to storing files. Any user on the network can store files on the server. A print server is a computer that manages one or more printers, and a network server is a computer that manages network traffic. A database server is a computer system that processes database queries. Servers are often dedicated, meaning that they perform no other tasks besides their server tasks. On multiprocessing operating systems, however, a single computer can execute several programs at once. A server in this case could refer to the program that is managing resources rather than the entire computer.
A patch panel is generally a panel of ports contained together that connects incoming and outgoing communication lines of a communication system. For example, in a LAN, the patch panel can connect the network's computers to each other and to the outside lines that enable the LAN to connect to the Internet or another WAN. Connections can generally be made with patch cords and the patch panel allows circuits to be arranged and rearranged by plugging and unplugging the patch cords.
The connecting of servers via patch panels to backend switches can simplify the network architecture and reduce latency in servers through cross switch traffic. It is generally considered to be cleaner and more efficient than other methods of connection. While smaller server farm implementations may require a single or a couple of switches to connect all the servers, larger server infrastructures require significantly more switches.
An exemplary case can be 240 servers, with each of these servers having two network interfaces. The 240 servers can be divided into ten server racks. Hence, each rack contains 24 servers. Each rack can have one 48-port switch that connects each of the servers on both ports. Each of these switches can be connected to a gigabit switch having at least 10 ports. For the purpose of this example, a common three-tiered architecture of servers can be utilized. This three-tiered architecture includes database servers that store information, application servers that perform complicated tasks, and front-end web servers that interface between the end user and the application or database servers. In this example, 24 of the 240 servers can be the database servers, 108 servers can be the application servers, and 108 servers can be the front-end web servers. Each rack can have two or three database servers, ten or eleven application servers, and ten or eleven web servers. The number of types of servers in each rack can vary widely.
A request can come in to a web server located on a first rack, where the request can pass from the server through a switch on the first rack to the gigabit switch. The switch latency to hear the request can be about 0.2 ms (milliseconds). The gigabit switch can then pass the request off to an application server located on a second rack and to a switch located on the second rack. Now, a switch latency of 0.4 ms is introduced by passing through the two switches. This can be repeated on the receive end of the transaction, from the application server back to the front end web server for a total request time to simply establish initial communication between servers of 0.8 ms. This effort can further be duplicated to communicate with the database server located on a third rack with a third switch.
As shown above, the number of hops from one server to another server is at least two—a first hop from the switch of the first server to the gigabit switch and a second hop from the gigabit switch to the second switch on which the target server is located. When this latency is multiplied by millions of requests, the lag in response can be increased by the millions. This latency can lead to slower response times and fewer requests being honored because the servers are constantly waiting for responses from one another. Eventually more machines must be implemented to compensate for the latency. Additionally, such latency from an end-user perspective (the end user being a user attempting to access the servers) results in slower load times, such as the loading of a web page in the prior example.
Patch panels can effectively reduce the number of switches in the system. Therefore, the number of hops can also be reduced. Instead of connecting to an intermediary switch at the top of each rack, a server can be connected to a patch panel that runs the server directly into a centralized switch. Using a large 480-port switch, all machines can be connected on both interfaces with half the latency and zero hops. Connections between servers can be handled directly on the switch and are thus more immediate.
Patch panels may be utilized in a wide variety of different systems. For example, patch panel can be utilized in a telecommunications closet, within the environs of a data center, or in other places. In certain embodiments, the patch panel can connect servers on a rack mount to switches or other network devices or any device utilizing a RJ-45 Ethernet jack.
An RJ-45 jack is a Registered Jack-45, which is an eight-wire connector used commonly to connect computers onto a local-area network (LAN), especially Ethernet. RJ-45 connectors look similar to the ubiquitous RJ-11 connectors used for connecting telephone equipment, but they are somewhat wider.
Vertical Patch Panel
In certain embodiments, the height L of vertical patch panel 100 can be about 72 inches (i.e., 6 feet) and the width W can be about five inches. In other embodiments, the height L of the vertical patch panel 100 can be at least 12 inches. The depth or thickness T of the vertical patch panel 100 can also vary. For example, the left edge as well as a section of the top and bottom of the panel can be curved. The depth or thickness T of the vertical patch panel along the curved edge can be about 7/16 inches. The thickness of the strip 116 in the center of the patch panel 100, which holds the various punch-down blocks 118, can be about 1/16 inch. Thus, the patch panel 100 can vary in thickness between the edges and the center.
The height L, width W, and depth T of vertical patch panel 100 can be varied as needed. For example, when vertical patch panel 100 is utilized in a home or office setting, the measurements may be adjusted as needed. For example, height L can be 12 inches, width W may be about 5 inches, and depth T of the curved edges can be about 7/16 of an inch while the thickness T on the opposing face (center) can be about 1/16 of an inch.
Moreover, curved edges on the top and bottom of panel 100 can extend 3.25 inches from the left edge of panel 100 on both the top and bottom (C1). On the right side of patch panel 100 are shown a series of 123 oval holes 120. The holes 120 in strip 116 can be arranged vertically in repeating sets of three, with center-to-center separations of 0.5 inch (12.7 mm), 0.625 inch (15.875 mm), 0.625 inch (15.875 mm). Therefore, the hole pattern can repeat every 1.75 inches (44.45 mm). The number of holes and the repetition patterns can vary within the scope of the invention, with the numbers set forth above being exemplary for one embodiment.
The holes 120 can represent 41U (that is, the height of 41 rack units). In a server rack having 42U, this can allow the vertical patch panel 100 to be mounted to the server rack with 1U or 1.75 inches remaining at either the top or bottom of the server rack.
The vertical patch panel 100 can be made from metal, although other materials can also be used. As described above, the vertical patch panel 100 has a height L that is substantially greater than the width W. During use, this allows the vertical patch panel 100 to be mounted on the side of a server rack and to service the entire server rack with cables of common length.
In certain embodiments, after connecting the required cables in the vertical patch panel 100, the vertical patch panel 100 can be mounted to one or more racks. For example, vertical patch panel 100 can be mounted via holes 120 that correspond to the spacing of a single rack unit. The vertical patch panel 100 can be mounted on the side of a rack. In addition, the vertical patch panel 100 can be mounted externally to the rack so that the vertical patch panel 100 does not take up any space in the interior of the rack.
Strip 116 can also have twelve rectangular holes 122 where the punch-down blocks 118 can be mounted (shown in
In certain embodiments, punch-down blocks 118 can contain about 8 ports, which may be an industry standard 8 port punch-down block. This can reduce overall cost of the invention as customers may select the type of punch down block 118 to be used. In certain embodiments, the ports can be RJ-45 jack ports and can be located in the front of vertical patch panel 100. The punch-down block 118 can use a variety of RJ-45 cables, including Category 5 (Cat5), Category 5-Enhanced (Cat5-E), or Category 6 (Cat6). In certain embodiments, vertical patch panel 100 can contain about 8×2×6=96 ports, but the total number of ports can be changed as needed. For example, in the embodiment of
Other utilizations of the patch panels include, but are not limited to, the ability to utilize the RJ-45 jacks for additional functionality, whether it is for keyboard, video, and mouse or KVM over Ethernet or to connect to management devices such as serial consoles or IPMI devices. The port density on the vertical patch panel 100 allows for many possibilities in configuration and usage. In one embodiment, the network ports are not precisely vertically oriented, but remain in a predominantly vertical orientation. Thus, the network ports need not be exactly vertically oriented, which permits variation in port density and alignment. Altering the orientation of the network ports can reduce the amount of cable necessary to connect devices to the patch panel.
The height H, width W, and depth T of vertical patch panel 100 can be varied as needed. For example, when vertical patch panel 100 is utilized in a home or office setting, the measurements may be adjusted as needed. For example, height H can be 12 inches, width W may be about 5 inches, and depth T of the curved edges can be about 7/16 inches while the thickness T on the opposing face (center) can be about 1/16 inch. The dimensions of W, L, and T can vary within the scope of the invention, with the numbers set forth above being exemplary for one embodiment.
Strip 116 can also have twelve rectangular holes 122 where the punch-down blocks 118 can be mounted (shown in
As shown, vertical patch panel 100 can further include cable guides 124. In this embodiment, each pair of punch down blocks 118 can be accompanied by a pair of cable guides 124. For example, as shown in
The server racks shown in
In a horizontal mount patch panel, the cable bundles connecting the patch panel to the patch panel on the server row are typically bundled over the rack and run on ladder racks to their destination.
The cable symbol 301 in
In the vertical patch panel 100 of the invention, the patch cables can be bundled, gradually building up to a top bundle that is then run to the bundle at the top. As this is mounted in a semi-permanent fashion, the installer does not need to run all the bundled wire; they may opt to just run the pairs of wires that are necessary without the cable shielding. This would reduce the size in back of the panel. Once mounted, the patch panel is intended to remain undisturbed until such a time as the owner no longer needs it. The patch panel may be upgraded by replacing the punch-down blocks. Since the punch-down blocks can be available as a standard part, in a standard form factor, this can be done without having to replace the shell of the patch panel.
The cable lengths for connecting servers 412a through 412c to horizontal patch panel 400 vary widely in length. The cable connecting the top server 412a to horizontal patch panel 400 is approximately 3 feet in length. However, the cable connecting server 412b to horizontal patch panel 400 can be about 5.5 feet in length, and the cable connecting server 412c to horizontal patch panel 400 can be about 9.5 feet in length. Thus, when a horizontal patch panel 400 is used with a server rack, cable lengths can vary widely.
Much like
Hence, by moving the patch panels to a vertical orientation, not only is space saved, but cabling is simplified. The ports can be brought closer to the servers and can eliminate the need for cables of differing lengths because all the cables are substantially equidistant from the ports of the vertical patch panel to the port on the server. This can reduce overall cost of cabling because shorter cables cost less money and the number of cables of the same type increases. This can simplify management in that only one type of cable needs to be stocked. Moreover, because the servers' cabling can look the same, it also has an added benefit of providing improved aesthetics. Additionally, if the network ports are not precisely vertically oriented, but have a slightly slanted orientation, the required cable lengths can be further standardized and the amount of cabling further reduced.
The number of servers on the racks 410 and 420, if fully stocked, can also differ. In rack 410 having horizontal patch panel 400, shown with no additional cable guides, no more than 38 servers can be installed in the illustrated embodiment. However, in rack 420 having vertical patch panel 100, 42 servers can be installed.
Moreover, conventional horizontal patch panel 400 is essentially composed of four 24 port horizontal strips and is designed to be mounted on two posts of server rack 410. However, vertical patch panel 100 can be designed to be mounted on just one post of rack server 420.
In addition, mounted cable spacers (through which cables can be run) were provided between servers (typically spaced apart at least about 5 inches) in the prior art. In the present invention, however, there is no need for running cables through these spacers because the cables can go directly to the devices. Hence, the spaces between the servers can be utilized to run the patch cables that connect the patch panels to the centralized patch panels where the switch or switches are connected.
In an embodiment, the vertical patch panel 500 also includes a hinge 501 oriented along the length L of the vertical patch panel 500. When the vertical patch panel 500 is mounted on the side of a server rack, or other rack, the hinge 501 allows adjustment of the vertical patch panel 500 relative to the server rack, or other rack, to be adjusted. This allows easier access to the vertical patch panel 500 while reducing reduce the area necessary to house the combination of server rack and the vertical patch panel 500. Hinge 501 can also reduce the amount of space necessary for the vertical patch panel 500 by allowing it to be adjusted so it protrudes less than its full width W from the server rack, or other rack. Additionally, hinge 501 allows for easier access to the rear of the vertical patch panel. In other embodiments, the vertical patch panel 500 includes multiple hinges oriented along the length L to adjust position relative to the server rack or other device. In alternative embodiments, the vertical patch panel 500 includes both a hinge 501 oriented along length L and a cable guide 124.
The vertical patch panels of the present invention reduce cable lengths, simplify design layouts, increase port density, provide enhanced aesthetics due to symmetry of cabling, and significantly reduce costs of ownership. Moreover, the present invention has the added benefit of providing a flexible design as different types of punch-down blocks (such as Cat5 or Cat6 ) can be specified at any time. For example, punch-down blocks can be replaced to any desired port types at any time.
Upon review of the description and embodiments of the present invention, those skilled in the art will understand that modifications and equivalent substitutions may be performed in carrying out the invention without departing from the scope of the invention. Thus, the invention is not meant to be limited by the embodiments described explicitly above, and is limited only by the claims that follow.
This application claims a benefit of, and priority under 35 USC § 119(e) to U.S. Provisional Patent Application No. 60/732,514, filed on Nov. 2, 2005, the content of which is hereby incorporated by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 60732514 | Nov 2005 | US |
