What Equipment do we use? - Page 3

Configuring a LAN party network - What Equipment do we use?

Details: Category: Networking; Published: Thursday, 01 December 2011 13:10

Page 3 of 3

What Equipment do we use?

Just like mentioned before, our equipment it broken down by our backbone and then our table switches. All of our switching equipment is Trendnet. Doing this makes support easier if you ever have problems and will cause fewer issues when hooking everything up. Networking is all about standards, but we have still found it to give us fewer issues when using only one brand over a mix and match of switches.

TEG-240WS c2_d1_2

Our Backbone Switch

We currently only need one backbone switch and that includes running all of the staff and our servers on that same switch to give the best connectivity to game servers hosted on staff PC’s. For that we use the Trendnet TEG-240WS 24 Port Gigabit Web Smart Switch. Here are the Specifications

Standards	IEEE 802.3 10Base-T IEEE 802.3u 100Base-TX IEEE 802.3ab 1000Base-T IEEE 802.3z 1000Base-SX/LX (Mini-GBIC) IEEE 802.3x Flow Control and Back Pressure IEEE 802.1D Spanning Tree Protocol IEEE 802.1p QoS IEEE 802.1Q VLAN Tag IEEE 802.1X Authentication SNMP v1
Protocol	CSMA/CD
Transmission Method	Store-and-Forward
Interface	24 x 10/100/1000Mbps Auto-MDIX RJ-45 ports 2 x 1000Base-SX/LX Mini-GBIC slots (shared with Gigabit ports 23-24)
Network Media	Ethernet: UTP/STP Cat. 3, 4, 5 up to 100m Fast Ethernet: UTP/STP Cat. 5, 5e up to 100m Gigabit: UTP/STP Cat. 5, 5e, 6 up to 100m
Data Transfer Rate	10Mbps: 10/20Mbps (Half/Full-Duplex) 100Mbps: 100/200Mbps (Half/Full-Duplex) 1000Mbps: 2000Mbps (Full-Duplex)
Data RAM Buffers	512KBytes per device
Filtering Address Table	8K entries per device
Switch Fabric	48Gbps forwarding capacity
Diagnostic LEDs	Per Unit: Power, System Per Copper Gigabit Port: Link/ACT, 1000M, 100M Per Mini-GBIC Port: Link/ACT, 1000M
Power Supply	100 ~ 240VAC 50/60Hz, internal universal switching power
Power Consumption	35 watts (max)
Dimension	440 x 210 x 44mm (17.3 x 8.3 x 1.73in.)
Weight	3kg (6.6lb)
Temperature	Operating: 0°C ~ 40°C (32°F ~ 104°F) Storage: -10°C to 70°C (14°F to 158°F)
Humidity	Operating: 10 % ~ 90 % (non-condensing) Storage: 5 % ~ 90 %(non-condensing)
Certifications	CE, FCC

As you can see the TEG-240WS has all of the requirements that we set forth before. It has the added benefit of giving us mini-GBIC slots for fiber if we ever need to expand beyond one backbone switch. Going with a Web Smart switch gives us most of the benefits of a managed switch (all of the features we need), while being a little cheaper in price. Buying managed Gigabit switches is still VERY expensive, even when used.

TEG S24Dg_d1_2

Table Switches

Our table switches are a lot simpler than our Backbone and are considerably cheaper. We went Trendnet 24 Port Gigabit GREENnet Switches Model number TEG-S24Dg. These fulfill all of our requirements including size and even have a full 48Gbps forwarding capacity, something that isn’t a requirement on table switches. Here are the specifications.

Standards	IEEE 802.3 10Base-T IEEE 802.3u 100Base-TX IEEE 802.3ab 1000Base-T IEEE 802.3x Flow Control IEEE 802.3az Energy Efficient Ethernet
Network Media	Ethernet: Cat. 5 up to 100 m Fast Ethernet: Cat. 5 and 5e up to 100 m Gigabit Ethernet: Cat. 5e and 6 up to 100 m
Data Rate	Ethernet: 10Mbps/20Mbps (half/full-duplex) Fast Ethernet: 100Mbps/200Mbps (half/full-duplex) Gigabit Ethernet: 2000Mbps (full-duplex)
Protocol / Topology	CSMA/CD, Star
Power Consumption	13 Watts (max.)
Switch Fabric	48Gbps
Interface	24 x 10/100/1000 Mbps Auto-MDIX RJ-45 ports
Power Button	On / off power button
Data RAM Buffer	3.5 MBits
Filtering Table	8 K entries
Jumbo Frame Support	Up to 9216Bytes
Diagnostic LED	Power, Link/ACT, 100Mbps, 1000Mbp
Power	100~240VAC, 50/60Hz Internal universal switching power
Dimension	280 x 180 x 44 mm (11 x 7 x 1.7 in.)
Weight	1.7 kg (3.8 lbs.)
Temperature	Operating: 0° ~ 40° C (32° ~ 104° F) Storage: -10° ~ 70° C (14° ~ 158° F)
Humidity	Max. 90% (non-condensing)
Certifications	CE, FCC

I mentioned power usage earlier, the GREENnet Switches are designed to only use minimum voltage required meaning we will see the most efficient use of power as possible. The MAX power usage is a whopping 13 watts, when compared to the backbones 35 watts it’s even more impressive. That’s of course at maximum; depending on connection lengths and usage power consumption can be even less (70% less according to Trendnet). Another benefit we have found in the TEG-S24Dg that we didn’t realize when ordering was the built in power switch. This is great for resetting the network; you don’t even have to unplug the switch.

All in all we have been extremely happy with the network. We have been using the full Gigabit network using the Trendnet switches for a few events now and we haven’t had one network related issue. The internet is a whole different issue. We can cover how to cover internet at your event in another write-up. Best of luck with your event and feel free to comment if you have any questions. If this write-up is helpful for setting up your event, please post on the comments here and let us know! We would love to hear about the events that have been helped.

TEG-S24Dg d3_2

Author Bio

Author: garfi3ldWebsite: http://lanoc.org

Editor-in-chief

You might call him obsessed or just a hardcore geek. Wes's obsession with gaming hardware and gadgets isn't anything new, he could be found taking things apart even as a child. When not poking around in PC's he can be found playing League of Legends, Awesomenauts, or Civilization 5 or watching a wide variety of TV shows and Movies. A car guy at heart, the same things that draw him into tweaking cars apply when building good looking fast computers. If you are interested in writing for Wes here at LanOC you can reach out to him directly using our contact form.

Posts in discussion: Configuring a LAN party network

garfi3ld replied the topic: #21258 01 Dec 2011 19:12

A peak into our LAN's network and a few times for those of you who want to run your own LAN's

Angrypirate replied the topic: #21259 01 Dec 2011 19:24

This is exactly what I've been looking for...

Angrypirate replied the topic: #21261 01 Dec 2011 19:44

I am very curious to know how all of the internet traffic is handled and what you use for load balancing, if needed.

garfi3ld replied the topic: #21262 01 Dec 2011 19:50

I was planning on doing a second writeup on internet and handing the internet traffic if this was well received

L0rdG1gabyt3 replied the topic: #21264 01 Dec 2011 20:50

Nice writeup. Thanks for the info! Looking forward to v10!

DrDeath replied the topic: #21295 02 Dec 2011 00:38

Good article man (read the whole thing). The only thing I would consider, is hanging all dedicated servers off their own switch, which is uplinked directly to the primary switch. The reason for this would be, if anyone is transferring anything from a server, it's usually a large file (i.e. gaming, movies or music). You don't want that traffic propagating the entire network, from one table to another. Instead, traffic is going from the server switch to the table switch that requested it. It's not a huge deal, and you're not going to see great benefits from it, but it's considered best practices.

garfi3ld replied the topic: #21297 02 Dec 2011 00:55

Unless I'm confused by your post, we already do that. The backbone switch is where all of the servers are hooked up along with the uplink from each table switch.

Angrypirate replied the topic: #24316 28 Mar 2012 04:35

So Wes, how's that followup on handling the internet traffic coming along?

Twodavez replied the topic: #24322 28 Mar 2012 05:46

Too bad you couldn't show the actual stats while at the LAN and condense it down to a nice 10-15 min video. It could be someting you could post on Youtube and even generate a little revenue for your efforts.

garfi3ld replied the topic: #24324 28 Mar 2012 05:49

Angrypirate wrote: So Wes, how's that followup on handling the internet traffic coming along?

we were waiting on the new router that we got for this event. But as you all noticed the new router had problems when we hit about 100 people. Once I get through all of our sponsorship followups I will be working with TP-Link to track down the problem.

Once we do that, we can finally do the writeup

garfi3ld replied the topic: #24325 28 Mar 2012 05:51

Twodavez wrote: Too bad you couldn't show the actual stats while at the LAN and condense it down to a nice 10-15 min video. It could be someting you could post on Youtube and even generate a little revenue for your efforts.

I really want to get setup and cleanup videos to show what goes into it all also. I have a feeling we will be following up with this specific writeup a few times until we have covered everything needed to know about setting up a LAN

Angrypirate replied the topic: #24375 29 Mar 2012 21:51

Good to hear. Can't wait to get all of your magical secrets...

garfi3ld replied the topic: #24377 29 Mar 2012 22:14

keep the feedback coming, I'm interested in what else you guys would like to see. To help me break up the writeups

Sideout replied the topic: #24469 04 Apr 2012 05:31

I think you should look at using PFSense for your router. Not only will you not reach a performance block with it but since it is open source , you can get some nice addons , bandwidthD is one , that will give you some interesting stats.

Here is the link (www.pfsense.org/) . You can run it on basically any type of PC you have and you can add in multiple NIC's for different interfaces.

I use it at work to provide a captive portal on our public WiFi network.

An x86 based router is going to be able to handle the higher PPS compared to a router like a TP-Link.

Back in the day (2004 / 2005), we ran PFsense on an Intel PII-400 with 512MB of RAM and 2 NIC's with a small hard drive and at one lan party we logged and passed 16 TBytes of LAN / WAN traffic. Additionally we had two cable connections that were load balance and QoS based on traffic type.

It might take a bit more to setup but it is a much better solution imo.

neokeelo replied the topic: #24581 10 Apr 2012 20:08

Nice article.

Im trying to help some buddies out with planning a LAN in a few months back in Tennessee.

If you wrote another article explaining how to manage the internet and also the name and settings of that LAN Check-In system you guys were using, it would be very helpful I think.

What is the name of that program you use to check everyone in and register for events?

From all of the events I have been to (probably over 50) LanOC was definitely the smoothest running event I have been to so far. Nice job guys.

Sideout replied the topic: #24610 12 Apr 2012 22:22

Not to give a spoiler alert but you might be able to see live LAN / WAN network stats at the next NeXus LAN event.

If the LanOC guys are there and are interested in doing something similar , I can help them out on setting it up.

Dreyvas replied the topic: #24611 12 Apr 2012 23:03

Sideout wrote: Not to give a spoiler alert but you might be able to see live LAN / WAN network stats at the next NeXus LAN event.

If the LanOC guys are there and are interested in doing something similar , I can help them out on setting it up.

Wes, I'm sure Alex and I could lend a hand with getting this set up as well if you are interested.

mianosm replied the topic: #25930 28 Jun 2012 20:47

Just surfing/browsing the web and stumbled upon this site.

Coming from another community that is actually doing the same thing as you guys - and just wanted to let you know the write up was great, and I look forward to any future installments for comparison and contrast to how we're currently setting things up.

The network for us seems absolutely solid, the issues we seem to run into more often than not is power issues (and mainly that comes about from the 4 guys with high performance quad-sli/xfire daisy chaining off of one outlet).

Keep up the great work!

CNO replied the topic: #36613 18 Apr 2015 08:15

Would it had been more advantageous to configure each switches ports bandwidth restrictions prior to deployment at an event? Granted, I haven't managed a LAN Party, however, the same considerations are at work in academic labs with 50+ workstations. We manage port traffic to limit web access, let alone LAN traffic, the same considerations are at work. Albeit, the majority of the students are 2-year IT students interfacing with mainly Cisco products. So, the decision not to configure the table switches, and configure the master switch, is a decision based on best practices pertaining to LAN Party experience? My thought here is without managing the local ports near the gamers, there is the potential for one or more clients to dominate a branch below the master, degrading neighboring clients. Were those ports also managed as well, balancing is insured, both at the master, as well, as each branch.

garfi3ld replied the topic: #36614 18 Apr 2015 13:51

Running managed switches on every table at a LAN event isn't really the best idea financially, at least early on. More to your point though, we haven't ever run into an issue where any one person or table has dominated the traffic the LAN traffic to the point where it caused slowdowns for people around them. File transfers up until very recently have always been limited by the transfer speed of the host and the write speed of the person pulling them, keeping any big transfers slow enough not to be a burden on users around them. With SSD's becoming more popular it is possible that we might run into the issue in the future, but like I said before the cost of buying managed switches for each table is huge and out of reach for most mid sized events.

That said, in a perfect situation setting a cap on each port at the table is a great way to prevent any one person from dominating network traffic for those around them

CNO replied the topic: #36615 18 Apr 2015 17:37

From the perspective of network management, I can see it is not a 'genre' consideration. I assume network monitoring played a role in your decision. ~~Do you have that data that supports this?~~ My interest in your decisions is looking at the statistics over the net, over the entire event. Armed with this info I might gain a greater understanding of the network traffic, bandwidth load, and node considerations.

Sideout replied the topic: #36617 18 Apr 2015 22:26

2. How much aggregate (inbound and outbond) traffic did we pass at LanOC v16?

a. WAN traffic - 1.6TB Aggregate
b. LAN traffic - 6.02TB Aggregate

This is a new LanOC record. We had 3 TWC 50/5 modems in a load balancing round robin config and most of the time , we were at a solid 35Mbits on all 3 modems with it peaking to 45Mbits at times.

This gives you an idea of how much traffic we were pushing at the event. We ran a Dell 2824 Switch at the core feeding into PFSense on a 1 gigabit connection. The Vmware was using a LAG group on the Vmware / Switch side at 4Gbit.

In a perfect world , yes you could get managed switches and limit bandwidth at that level. I dont see any sponsor ponying up that hardware though.

Another factor we have is that with multiple modems , it is not possible to get a true load balanced solution so while we have 3 modems to the Internet , it is not a true 150Mbit. Again having one pipe to manage for that would make things simple as well but that is not the hand we have been dealt.

The goal with any QoS / Filtering / Throttling should be to balance the game play with the users ability to download or surf as needed. This last LanOC , it was the best one we had from that aspect. It is not an exact science when you factor in 140+ people in the room running all kinds of stuff like Spotify , Hamachi , Bit Torrent , UTorrent , and others on their PC - all leeching for the best amount of bandwidth they can get. Factor in as well game updates the day of and other download needs and you have a large amount of demand for a small pipe.

Not to mention we had one modem pretty much dedicated to streaming the LAN live so that cut into the bandwidth we had available.

if your interested in my setup I used for PFSense you can find it in the forums here or over at the pfsense forums.