What is a patch panel? How it is different from (layer 2) switch?
EDIT: I think my primary question (and confusion) is this:
Why don't network connections go straight to the switch? Don't we still require 1-to-1 connections from switch to patch panel?
splattne has covered what a patch panel is, and why it's different to a switch.
To answer the last part of your question: the reason that host network connections don't go direct to switches is generally to do with ease of management. For example, desk locations on an office floor can be cabled back to a wiring closet patch panel which is labeled with the locations. You can then connect short patches ('tails' or 'whips') between the patch panel and the switch. This makes re-patching desk locations (for user moves etc) much simpler, as the desk->patch panel runs don't need to be touched at all.
In a data centre, a similar argument applies. If a server needs to be moved to a different subnet that is on a different physical switch to the one it's connected to, having intermediate patch panels is very useful. For example, many server rooms have an MDF (master distribution frame); all servers and all switch ports are cabled back to labeled patch panels on this frame. Then, creating a connection between a server and a switch is a simple case of a patch between two ports on the frame, rather than needing to have floor tiles lifted to run a new end-to-end patch.
EDIT: To add a few sample cabling topologies:
1) User floors.
[host]<<--patch-->>[floor port]<<--structured cabling-->>[wiring closet patch panel]<<--harnessed/bundled cabling-->>[wiring closet access switch]
2) Data centres, centralised access.
[host]<<--patch-->>[cabinet patch panel]<<--structured cabling-->>[master frame patch panel A]<<--patch-->>[master frame patch panel B]<<--harnessed/bundled cabling-->>[data centre access switch]
Note in the above, you could have another cabinet patch panel in the switch cabinet; however when using large modular switches (240+ ports per chassis), providing that many patch panel ports tends to use up valuable U-space in the cabinet; hence why these connections are often directly harnessed back to the master frame.
3) Data centres, distributed access (end of row).
[host]<<--patch-->>[cabinet patch panel]<<--harnessed/bundled cabling-->>[end of row access switch]
This kind of topology is often used with blade deployments, as the number of blade chassis you have deployed dictates precisely the number of ports you need to provision. Note the reduced physical flexibility, however - hosts must be cabled to switches in the same row. Your logical network design should take this into account.
4) Data centres, distributed access (top of rack).
[host]<<--patch-->>[top of rack access switch]
Potentially useful where you have a very homogeneous datacentre with lots of nodes with identical requirements.
Note these are just some examples - there are plenty of other approaches as well.
This is a patch panel:
From Wikipedia:
A patch panel or patch bay is a panel, typically rackmounted, that houses cable connections. One typically shorter patch cable will plug into the front side, whereas the back holds the connection of a much longer and more permanent cable.
The assembly of hardware is arranged so that a number of circuits, usually of the same or similar type, appear on jacks for monitoring, interconnecting, and testing circuits in a convenient, flexible manner.
Regarding your question about what's the difference between a patch panel and a switch: a patch panel has no "intelligence." It's simply used for connecting cables.
The counterpart to a network jack!
If you have big networks, you use CAT5e & CAT7 wires which are thicker than patch wires. From the Jack to your notebook, you've to use a patch wire.
To connect every network cable to the core switch which comes from the network jacks you need a big "panel" to patch the wires to the switches. That's the patch panel.
The inner life of a patch panel is similar to a network jack:
Patch panels are there to deal with big masses of cables!
The reason we use "patch panels" instead of terminating cables into ends is that a patch panel supports the cabling coming into the server room, protects the terminations from being knocked loose -- and terminating into a patch panel is extremely fast and easy. The punchdown tool used makes it a trivial task taking ~10 seconds per cable end to terminate, whereas putting ends on each cable and labeling them would take much longer.
The patch panel also helps keep the bundle of cables all neat coming into the closet. It is much easier to manage a number of 1.5' patch cables going between to firmly mounted units than it is to wield a stack of 50 cables.
Its also the way the telco does it historically.
I realize that there are lots of answers to this question already, but I have a simple way of looking this:
Think of a wall jack as a mini-patch panel. You wouldn't run a line from your switch directly to a particular PC, right? Well, you wouldn't run a line from a wall jack directly to a switch.
If you did run a line directly to a PC, you'd run into the inevitable problems like "what happens when the person wants to move their computer three feet and the line isn't long enough?"
Likewise, if you run a line directly to a switch, what happens when you need to plug that same line into a different switch in a nearby rack? Dang, the cable is an inch short!
So, any run through a building should be terminated at both ends... at patch panels and / or wall jacks. These runs are used for covering distance, and should not limit how you arrange the hardware that will eventually jack into these lines.
There are other benefits such as simplified labeling, connecting different cable types (solid to stranded), switch wear and tear, etc, but really a patch panel is just a nice way to terminate a bunch of lines that need to get from some other places to "around here somewhere". You'll see the same in plain old telephone systems: every phone jack on the wall is running to a punch block (same purpose as a patch panel) that's mounted somewhere near the telephone switch.
I didn't see this answer, I apologize if it's a dupe, but it seems to me that a big reason for using a patch panel in an office is because you often don't have switchports (which are expensive) to support every office drop (which are cheap). So, when you build out an office, you run some amount of drops to every office. For instance, in our building we had four drops in every office. At the time of construction (before drop ceilings are on and it's still easy) you run all of those to wire closets, and each cable goes to the back of a patch panel. In initial mode, you may only have enough switchports to connect one port in each office/cube. That might run a voip phone with a PC plugged into the back of it. As time moves on, maybe two people share an office. No problem, just connect a second patch panel port for that office to the switch. Running low on switch ports? Put a new switch in...
In other words it allows you to build lots of switch ports when the building is built or during a single painful recabling without necessarily needing switch ports for every run. Technically you could do this with just a bundle of cable ends I guess, but it's a heck of a lot easier and less likely to become unmanageable with a well labeled patch panel.
Oh, I just remembered another scenario. I once had a piece of equipment deployed in an office that required a crossover connection. The switch it was attached to wasn't smart enough to be configured to change over to xover mode, so I was able to run an xover cable (extremely well labeled as being such, of course) between the patch panel and the switch to get the device up. Now, technically I could have built a jumper and used a connector, or even re-ended the long cable run, but using jumpers sure is a lot easier and less error prone.
A short answer is to ease the transition from runs of cabling to the server or switch connection itself. You could crimp connectors on the ends of cables that you install, but it's easier to install and make changes and makes a neater installation if you punch those cables down on a patch panel, then make the final connection from there to the server using a patch cable. Also, the patch panel carries the burden of plugging and unplugging of cables. This protects the more expensive switch from wear and tear and damage.
Cables running through the walls and ceilings (etc.) of the building should be solid core (running from jack to jack), and then from the jack to the device a patch cable with stranded wires should be used.
You don't want solid core right to the device due to the lack of flexibility (solid core in a crimped connector tends to disconnect with minimal flexing), and trying to punch-down stranded is often like shooting pool with a rope.
The point of a patch panel is to prevent having to have dozens or hundreds of wall jacks in the server/network room. ;)
ANSI/EIA/TIA-568 is the Commercial Building Telecommunications Cabling Standard wherein a patch panel is defined as a connecting hardware system that facilitates cable termination and cabling administration using patch cords. It is important to note that connecting hardware has a limited life-cycle resulting from wear caused by multiple mating cycles. Whereas a Cisco switch could cost in excess of $30-40K a patch panel cost a couple hundred dollars. For example it is not difficult to imagine that at some future date someone yanks cables in the ceiling plenum which are directly connected to the switch instead of patch panel - such a scenario could damage costly equipment and possibly result in unacceptable downtime for network users.
Additionally field test manufacturers test to the above referenced standard. A "link test" requires that a tester be plugged into a communications outlet (i.e., data port) on the near side and a patch panel on the far end (or visa versa). In other words you could only test for continuity, wire mapping and shorts but you could not get any verifiable data as to the performance of the link based on industry standard test; making it difficult to diagnose network performance issues.
Let us redefine progress to mean that just because we can do a thing (cf.; run a cord from the NIC card of a PC directly to the port on a switch), it does not necessarily mean we should do that thing.
I don't think this has been covered yet (it was covered in part by Captain Segfault's answer), but another common use for a patch panel is in fact, not to patch connect a network socket to a switch at all.
More than once I've seen a patch panel socket plugged into another patch panel socket. For whatever particular reason they needed to create an end-to-end direct-connection between two network sockets at opposite ends of the building. If you were cabled directly into the switch, this would be far more difficult to achieve.
Cat-5 can also be used for phone systems, where the plug into a custom PBX, not a network switch.
I think to address your confusion and doubt over the utilization of patch panels, consider these scenarios:
A) a data centre where the core switches are located a corner. to provide network connectivity to all servers and machines in the data centre, lengthy cables for each and every one of them has to be measured and cut and crimped and pulled from wherever they are positioned in the data centre over to the core switches. That sprawling sea of ethernet cables underneath the floorboard is not going to be a clean one....
B) an office space spanning five storeys in a building. The network switches are only located in the server room on the first storey. To connect everybody's workstations or network printer, they have to measure and cut and crimp and pull across the levels to reach switches. I don't want to work as the "network guy" in this office.
The patch panel provides the facility of "localisation". Dedicate a fixed number of switch ports to a panel, and this panel is meant to connect nodes that are located in a particular location. The cabling from switch to patch panel becomes fixed and permanent, can be neatly laid out, thus better understood and managed. Servers or workstations at that location connect to the patch panel -or even their own work desk port which extends from the patch panel - using shorter and easier-to-manage ethernet cables. It simply makes the distribution of network wiring much much much less messier and chaotic.
http://howtocable.com/dataCabling/
The choice is yours.
Something I haven't seen anyone address is price. Imagine this scenario.
8 racks, 40 machines per rack, one core switch
Rather than pay for two 24 port patch panels per rack, 16U of patch panel space next to your core switch, and then 320 GigE ports on your switch I found it cheaper to use a 48 port switch per rack and then run two GigE (or 10G) back to the core switch. You might need to be careful not to oversubscribe a rack switch as well as make sure entire farms aren't in a single rack in case you lose that switch.
As always the price varies over time so make sure you price both options, but in a datacenter it's almost always easier and cheaper to go with the rack switch to core switch topology.
The same question could be asked about say wall outlets for electricity - why do you have outlets in your house? Why not just run and plug the cable directly into the toaster/fridge/tv/lamp?
(ok somewhat bad example as an outlet would correspond to a client jack - but the reason for client jacks are the same as patch panels)
Patch panels can be very helpful for equipment that moves around a lot, or for more unusual connections.
As an example I've used them for connecting serial ports (over CAT-5 wiring) to serial servers; since patch panels are at the wire level that works fine, but if you just connected everything to switches that wouldn't work. I've also used them to connect analyzers and other such equipment, where the more permanent hardware was connected to switches by organized cable bundles.
