Difference between revisions of "Infrastructure:Network Architecture"

From the Linux and Unix Users Group at Virginia Teck Wiki
Jump to: navigation, search
imported>Pew
imported>Pew
(Historic)
Line 25: Line 25:
 
[http://no-www.org/ www. is deprecated.]
 
[http://no-www.org/ www. is deprecated.]
  
== Historic ==
+
See [[Deprecated Network]] for historic information
=== ECE Server Closet ===
 
==== Limitations ====
 
We are behind the ECE Whittemore NAT, which is on a single 100 Mbps CNS port. We have the following limitations:
 
* All adjustments to ECE DNS must be made through [mailto:rbrand7@vt.edu Brandon Russell]
 
* IP addresses are difficult to claim, because they must be forwarded through the NAT
 
* IPv6 is not supported behind the Whittemore NAT
 
 
 
Consequently, We must:
 
* Use an IPv6 tunnel if we want access to IPv6 addresses
 
* Keep all internal services (like NFS) on an internal network
 
<!--* Only one MAC address may appear on the port at a time (port security)
 
* There is no prefix delegation for IPv6, so each address must be individually requested via NDP.
 
 
 
This means we must:
 
* Use ARP proxying or 1-to-1 NAT for IPv4
 
* Use an NDP proxy for IPv6-->
 
 
 
==== Desired Setup ====
 
This is what I'm hoping to migrate us to:
 
* OpenWrt ([https://github.com/sbyx/odhcpd odhcpd] has built-in NDP proxying) or pfSense Router
 
** Partial: pfSense provides NATing on [[Infrastructure:Cyberdelia|cyberdelia]]
 
* An internal network smaller than a /8 (room for expansion)
 
** Done: 10.99.0.0/16
 
* IPsec (point-to-point and road warrior for users)
 
** Can be done through openWRT or pfSense
 
* Each VM host has a bridged ethernet port with a global IPv4 address and performs NAT to its VMs. Additional IPv4s are assigned as VMs as needed (e.g. milton and acidburn probably need their own)
 
** Done on [[Infrastructure:Cyberdelia|cyberdelia]]
 
* All internal IPv4 addresses are static leases assigned by [[Infrastructure:temp88191|the router]] or set statically '''and documented somewhere'''; hypervisors do not have their own networks unnecessarily like wood currently does.
 
** Internal network on [[Infrastructure:Cyberdelia|cyberdelia]] has static IPs or long-term leases.
 
** Cyberdelia still has too many internal networks, most of which are unnecessary.
 
* Each device has a global IPv6 address
 
** Currently provided through tunnel
 
 
 
 
 
=== CVL setup (deprecated) ===
 
 
 
Hardware:
 
* "luugtemp" or "temp88191": a Poweredge 2650 with 2 NICs configured as an Ubuntu router
 
* 8-port Gigabit unmanaged switch
 
* 48-port 100 Mbps managed switch (attached to sunway)
 
 
 
Port security evasion:
 
* A bash script named "Nat" which presumably does 1-to-1 NAT
 
* NDP proxying via https://npd6.github.io/npd6/
 
** This is broken an misconfigured. It doesn't properly add routes.
 
 
 
IPs / networks:
 
* temp88191 is 10.0.0.1/8 and 128.173.88.191. It provides DHCP on our internal interface
 
* Sunway has static IPs setup (10.0.97.10 to 10.0.97.28)
 
* Rackable servers: joey (10.0.4.10) and phantomphreak (10.0.4.11)
 
* cyberdelia's IPv4 is luug0.ece.vt.edu
 
** Port 9001 <-> 10.0.1.3 (cerealkiller)
 
** Port 9030 <-> 10.0.1.3 (cerealkiller)
 
* wood's IPv4 is luug1.ece.vt.edu
 
* milton's IPv4 is luug2.ece.vt.edu
 
* luug3.ece.vt.edu is (in theory) used by westinghouse (sunway's head node)
 
* acidburn's IPv4 is luug.ece.vt.edu
 
* acidburn has iodine configured as a DNS tunnel (10.152.78.1/27)
 
* Other tenants of our router: mjh.ece.vt.edu and mirror.ece.vt.edu
 
* 10.99.0.2/24 appears to be statically assigned to wood's guests.
 
 
 
Cyberdelia VMs - assigned 10.0.1.1/24 (not actually a separate subnet):
 
* dhcp-host=52:54:00:14:df:c2,10.0.1.1 # "mail" (not yet configured)
 
* dhcp-host=52:54:00:68:81:33,10.0.1.2 # crashoverride 2.0
 
* dhcp-host=52:54:00:40:9a:55,10.0.1.3 # Cerealkiller 2.0
 
  
 
[[Category:Infrastructure]]
 
[[Category:Infrastructure]]
 
[[Category:Needs restoration]]
 
[[Category:Needs restoration]]

Revision as of 01:47, 4 January 2018

TODO: change name to "Network Architecture" and change content to more generally describe how the univerersity network works, with specifics about VTLUUG's setup. Direct to Infrastructure:Network with talkinga about specific IPs and mention that link a lot throughout the page when we fix the wiki so this isn't confused Infrastructure:Network and move Historic section to and leave a link here

This is an attempt to document VTLUUG's overly complex networking setup. Apologies for the disorganization, this is mainly just a way to get everything in one place. --Mjh (talk) 21:43, 28 December 2014 (EST

Note: This is extraordinarily dated. Revisions are in progress, but currently, do not consider it to be remotely correct. --echarlie

Current

We currently have a ~1Gbit NI&S port in the ECE server attic

Hardware:

  • "luug5" or "temp88191": a Poweredge 2650 with 2 NICs configured as an Ubuntu 14.04 router
  • cyberdelia

Cyb has a private network for NFS on 10.99.0.0/24, and temp88191 does NDP proxying, static ARP using jkh's Nat script (see github), and hands out dhcp leases somewhere in 10.0.0.0/8

We have no other hardware in use

DNS

See also: Infrastructure:Network#DNS

www. Required for Subdomains

Requiring www. in front of many subdomains is a known but ignored issue. Apparently, nobody at CNS has bothered to fix it. Rumor has it that the utility of their 25-year-old maintenance scripts is to blame for this issue and other DNS complications like difficulty maintaining MX records. Resolving this issue must be done individually at each department by addressing the problem to the relevant network liasion.

www. is deprecated.

See Deprecated Network for historic information