Question about feasibility of proposed network
-
I'm involved in starting a new charter school. We're moving into a building that housed a public school and has existing network switches and a single gigabit ethernet cable running to each classroom. We plan to put a switch in each classroom with one port to the teacher's workstation, one port to an wireless AP and one or more ports to student work stations. The front offices will have switches connected to a workstation and a voip phone. I am planning to have separate vlans for each group, students, wireless, voip and teachers/staff. The wireless APs will be used to support Chromebooks. Evenually there will be 500 or so students each with a chrome book.
Does this sound reasonable? The existing switches are Dells but I don't have the model number handy. Also I want to be sure pfSense will support such a system.
I've never had to deal with a network this complex before.
Thanks,
Jim. -
while there technically doesn't seem an issue with this, it might be cheaper & easier to manage to pull some more cables (no clue how difficult this is, in your situation).
( i do school-IT for a living & know about the limited funds)
small managed switches still cost $80-$350 / piece. each need to be configured with their vlans. each need some form of cabinet, to physically protect them from tampering.i don't know the layout of the school in your situation, nor the total amount of end points.
personally i'd put a small 10-12U cabinet in each (big)hallway/floor and pull wires towards the required endpoints. (with the exception of classes with say 20-30 endpoints, that justifies their own cabinet)
Put in 1 or 2 bigger switches (=24-48p) in the cabinets. The cabinets themselfs will have uplinks towards your central network/server rack.the latest 802.11ac-AP's needs 2 copper uplinks to fill their maximum theoretical throughput. who knows what'll happen in 2 years.
also, is the wiring in the building capable of moving to 10Gbe in a couple of years? Cat6a or better will be required.its very difficult to do major changes to a network when the school is operational. (teachers dislike people with a drill hammer while class is in session)
-
The wireless APs will be used to support Chromebooks. Evenually there will be 500 or so students each with a chrome book.
Does this sound reasonable?
I don't do school networks so I am not an authority on this. However, wireless networks are not switched networks, they behave a lot like old fashioned ordinary Ethernet where there is contention for the network. The more transmitting stations per segment, the slower it gets for everyone using it.
Lets say for example that due to the physical construction of the building that 802.11g bandwidth of 54Mbps is the maximum performance that is reliably achievable and that within the 2.4Ghz space you can get 3 bands operating without interference. That would give a theoretical maximum of 3 x 54Mbps = 162Mbps. Divided equally between 500 wireless chromebooks = 0.324Mbps (324Kbps each). Lets also assume that you can throw in some 5Ghz kit, and due to the size of the campus run 6 2.4Ghz bands without interference. So being generous, ignoring any degradation of wireless performance and assuming you can achieve four times what I have estimated as a 'worst' case, 4 x 162Mbps = 648Mbps, divided by 500 chromebooks = 1.3Mbps each.
I think you need to discuss this with someone that has practical experience of large scale wireless network deployment. It will probably need more than one wireless survey to get access point locations determined for the best performance but without specialist advice you risk spending good money on something that performs no better than individual dial-up modem access per student.
Does this sound reasonable?
-
while there technically doesn't seem an issue with this, it might be cheaper & easier to manage to pull some more cables (no clue how difficult this is, in your situation).
( i do school-IT for a living & know about the limited funds)
small managed switches still cost $80-$350 / piece. each need to be configured with their vlans. each need some form of cabinet, to physically protect them from tampering.You already guessed it. It will be rather expensive to run additional cables and we're looking at donated hardware for the switches and APs. We have almost no budget. Fortunately our donor has access to a lot of recycled equipment.
Thanks,
Jim. -
I don't do school networks so I am not an authority on this. However, wireless networks are not switched networks, they behave a lot like old fashioned ordinary Ethernet where there is contention for the network. The more transmitting stations per segment, the slower it gets for everyone using it.
Lets say for example that due to the physical construction of the building that 802.11g bandwidth of 54Mbps is the maximum performance that is reliably achievable and that within the 2.4Ghz space you can get 3 bands operating without interference. That would give a theoretical maximum of 3 x 54Mbps = 162Mbps. Divided equally between 500 wireless chromebooks = 0.324Mbps (324Kbps each). Lets also assume that you can throw in some 5Ghz kit, and due to the size of the campus run 6 2.4Ghz bands without interference. So being generous, ignoring any degradation of wireless performance and assuming you can achieve four times what I have estimated as a 'worst' case, 4 x 162Mbps = 648Mbps, divided by 500 chromebooks = 1.3Mbps each.
I think you need to discuss this with someone that has practical experience of large scale wireless network deployment. It will probably need more than one wireless survey to get access point locations determined for the best performance but without specialist advice you risk spending good money on something that performs no better than individual dial-up modem access per student.
Does this sound reasonable?
I hear what you are saying, but for the next year we have a very small budget. At least initially we'll only have about 100 chrome books to worry about so I'm hoping that we will be able to manage for a year with a single AP in each room. I think a survey will consist of placing APs in various places and see what kind of coverage we have, rather than a formal survey which costs money.
Thank you for the analysis. It's very helpful. Next year we'll have more funding and may be able to afford a specialist.
Jim.
-
Designing a wireless network for 16 classrooms with 30 laptops in each and a cafeteria with 500 laptops in it are two very different things.
I don't see pfSense as being a limiting factor in anything you are trying to do. Your challenges are at layers 1 and 2 - how do you run the necessary cable and how to you configure/secure the switch ports.
One gig-e drop to each classroom with a managed switch is going to be enough if you can secure it.
-
I think a survey will consist of placing APs in various places and see what kind of coverage we have, rather than a formal survey which costs money.
By learning how to do a survey properly yourself you will learn how to spot advise that as more likely to sell you more kit than just enough kit to get the job done.
http://www.wi-fiplanet.com/tutorials/article.php/3761356/How-to-Conduct-a-Wireless-Site-Survey.htm
http://www.networkworld.com/article/2925081/wi-fi/7-free-wi-fi-stumbling-and-surveying-tools-for-windows-and-mac.htmlYou should try out Kismet https://www.kismetwireless.net/. Data collection will be manual but it will only cost you your time and any knowledge gained using Kismet will be useful if you choose to use it as part of your security system later.
-
Not mentioned in that link is NetSpot for the Mac. You can conduct small surveys with the free version. http://www.netspotapp.com/
-
Designing a wireless network for 16 classrooms with 30 laptops in each and a cafeteria with 500 laptops in it are two very different things.
I don't see pfSense as being a limiting factor in anything you are trying to do. Your challenges are at layers 1 and 2 - how do you run the necessary cable and how to you configure/secure the switch ports.
One gig-e drop to each classroom with a managed switch is going to be enough if you can secure it.
Thanks. It's going to be a challenge. I don't have any formal education in network design/administration and have never ventured into enterprise wide configurations, so I'm really going to have to do some quick learning. Since it's a school we've got to filter various sites. I'm looking at squid for that. Do squid and pfsense play well together? Excuse me if that's a really dumb question. I'm venturing into unknown territory.
Jim.
-
I think a survey will consist of placing APs in various places and see what kind of coverage we have, rather than a formal survey which costs money.
By learning how to do a survey properly yourself you will learn how to spot advise that as more likely to sell you more kit than just enough kit to get the job done.
http://www.wi-fiplanet.com/tutorials/article.php/3761356/How-to-Conduct-a-Wireless-Site-Survey.htm
http://www.networkworld.com/article/2925081/wi-fi/7-free-wi-fi-stumbling-and-surveying-tools-for-windows-and-mac.htmlYou should try out Kismet https://www.kismetwireless.net/. Data collection will be manual but it will only cost you your time and any knowledge gained using Kismet will be useful if you choose to use it as part of your security system later.
That's great information. This will be an exclusively Linux site but I may be able to shake a Windows laptop loose to do a survey. I may start with Kismet since it runs on Linux.
Thanks,
Jim. -
Not mentioned in that link is NetSpot for the Mac. You can conduct small surveys with the free version. http://www.netspotapp.com/
Thanks, but I have no access to Apple hardware.
Jim
-
This post is deleted!
