Cisco Firewall :: 1811 / Zone Based FW With Non-standard HTTP Port
Apr 4, 2011
We are testing a Zone Based FW config since 1month, everything run smooth but we're having problem ( big slow speed access ) when a user try to reach a website on a non-standard port ( 8080 in that case ). All the trafic stay in our LAN, using a IPSEC/EZVPN connection between the 2 sites.As soon as I have disabled the Zone Based FW, the speed was much better.
I'm sure I'm missing a parameter to fix that problem but I tried many different options and I didn't find anything yet. All the routers are Cisco 1811 running adv IP Services 15.1.2.T1 IOS.A port-map has been created to map the port 8080 to the HTTP protocol for the inspection.The PC will have an IP address in the 10.2.2.x/24 and will access a server on 10.2.3.x/24, both devices are part of the zone private in each site/LAN.All the access between sites are managed by an ASA; the IPSEC/EZVPN peer.Little summary, it's gonna be something like : SiteA with a PC on private zone then on public zone for the EZVPN to SiteB on public zone and then private zone to access the server in the LAN.
I have two 1811's connected in a lab using a ipsec vpn tunnel (using a switch to simulate an internet connection between them).I am trying to configure one of the routers as a ZBPF just to allow a remote windows login (DC on the firewalled side, workstations on the other side).I'm trying to verify that the zbpf is working, but it doesn't seem to stop anything. I had match icmp added to the class-map, but took it out to test if icmp would fail. It didn't. Basically, I don't think the firewall is working at all. Any thoughts on how I can configure this so that the policies will work between zone-pairs?
Here's an quick drawing:
Here are the configurations:
Local router: hostname sdc-1811-LocalLab ! boot-start-marker boot-end-marker ! no aaa new-model ! resource policy
I'm having a few problems at the moment with a zone based firewall setup. The more I looked into the problems the more I question whether I need the ZBF or not.My network is pretty simple. 1 Internet connection and 1 LAN interface and a few site to site vpns to the router.So what do people think to having this kind of set up and not using a ZBF?
We had a problem with SMTP inspection dropping some regular emails (Cisco 2901 IOS 15.0).Incoming mails are going thru Spam and Virus Blocker so that bypassing SMTP inspection is not security issue in this case.
I've been trying to configured Websense urlfiltering using ZFW feature on my Cisco 881G router. The router is running on IOS 15.0(1)M with Advanced IP Services. And I have confirmed it supports urlfilter feature.
This is what I tried to accomplish but IOS version 15.0x seems to have different command set. ----------------------- class-map type inspect httptraffic match protocol http parameter-map type urlfilter param server vendor websense 10.20.30.40 [Code]...
I am getting ready to deploy a 3945 ISR to serve as an internet and core router for and remote site. I will be terminating a site-to-site VPN tunnel on it and also configuring a zone based firewall config between my "outside" (internet link) and "inside" (all internal nets). My question is about how to approach securing the WAN interface with the Zone based FW in place?what kind of ACL do I need beyond those allowing and restricting remote access to the outside ip?
I have a cisco 1841 router , and i want to configure zone based firewall on it. But the document of zone based firewall only said that "after 12.4(6)T" can support zone based firewall. I use the ios " c1841-ipbasek9-mz.124-15.T9.bin ", but it can't support ZFW. What kind of ios support ZFW. for example: ipbase, ent base, ip service ,advent etc.
we are experiencing performance issues on ASR 1004 with ZBF as our campus edge router.Symptoms:
- sending small packets from inside zone to outside zone, for example UDP packets without payload - this way I can generate up to 150.000 pps traffic (testing with packeth software, but we have had a real example with some kind of worm/virus) - CPU load is about 1% (yes one!) to 2% all time !! (weird) - ASR response to pings rises very quickly up to 5 seconds which makes box unusable dropping everything what goes through ZBF (so internet connection is gone) - if I do the ping directly from box, it seems to work fine (no rules from self to outside zone in ZBF) - if I remove interfaces from inside and outside zone (so disabling ZBF) and do the test again, ASR response goes from normal (0.2ms) up to 2ms (still sending 150.000 pps) and everything seems to work fine)
According to Cisco Datasheets: routing, Qos, Zbf ... on ASR 1000 with RP1, ESP10 should be done in hardware with up to 17.000.000 pps performance.
I am looking to implement Zone-Based Firewall on some 2900 series routers (2911 and 2921.) Based on some research I've done it looks like the cisco2911-sec/k9 and cisco2921-sec/k9 bundles should be all I need. Is this correct, or is there some other licensing component that needs to be enabled for me to implement Zone-Based Firewall?
I am confiuring ZFW on a Cisco 2951 Router. The router has the following interfaces: [code]Port Channel 1, 1.5, 1.10, 1.15, 1.20 have been added to the zone called IN-OUT. All the subinterfaces correspond to an internal VLAN.The router is connected to a MPLS network and has a BGP peer on interface MPPP. Over the MPLS network, an ecrypted DMVPN tunnel to HQ has been built (tunnel 0). EIGRP is the routing protocol running over the tunnel.Traffic coming in from HQ has to be firewalled on this router (don't ask me why!!). As a result, I am configuring ZFW on this router.
1-The router itself does not need to be protected, only the servers in the remote offices. That being said, I am not planning to create any self zone on this router. I don't want to break BGP, therefore the MPPP interface will NOT belong to any zone. Is this the correct way to do it?
2-The tunnel 0 interface will belong to OUT-IN zone that will protect all incoming traffic into this site from HQ. So when writing class-maps for the traffic coming INTO this site, do I need to write any class-maps for EIGRP or ESP? My guess is no, since that traffic will not be coming into the site, but rather just terminating on the router.
I'm having an issue accessing a clients router on the WAN interface with Cisco config pro. I can get CLI access with SSH without any issue. I have port 22 and 443 allowed as management access from my public IP - SSH working fine but config pro being refused connection, Possibly a certificate issue?
we have an ASR1002 running zone-based-firewall with 2 zones:
I have a common ZFW-configuration on that interfaces, e.g.
<code> class-map type inspect match-any pass_cmap_in match access-group name pass-ipv4-in ! class-map type inspect match-any ph_cmap_in match access-group name ph-ipv4-in
There is some basic stuff in the Access-Lists; direction ph-ipv4-in contains basically "permit ip any any" and ph-ipv4-out contains some permits for certain services, but nothing else. The pass-ipv4-in/out ACL contains particularly the udp-500/4500-stuff as well as gre/esp/ah.
The xconnect is only built up correctly when I configure the interface in the zone_outside. The destination for the xconnect is an ASR9k. If I do not configure the zone on the L2VPN-Interface, only arp-packet are allowed to tgo through the tunnel.
The L2VPN connects a branch office to the network of "PH". Now the trouble starts: when they are putting a host in the branch office, DHCP via the L2VPn works fine, they can ping anything from the branch office-PC in their local network and reach all internal servers etc.
BUT if they want to go to a destination outside their network, it will not work properly. For example, the branch-office-PC can ping 18.104.22.168 fine, but when they try to connect to a website, e.g. www.google.com, they run into a timeout. Netstat says, that the http-syn is sent, but no ack is received.
whereas x.y.225.250 is the PC connected via L2VPN in the branch office to their local lan. When they put the same machine in their local lan directly behind the router (without l2vpn) everything works fine. When I switch off the firewall on the Gi0/0/0-Interface, the PC from the branch office also reaches its destination, so for me it looks like the firewall inspects the traffic going via Gi0/0/1 and L2VPN, what in my opinion, it should not do....
I'm trying to configure a zone-based firewall on an SR520 and am confused about the 'not' criterion. The 'zone-design-guide' says (my stress): Class- maps define the traffic that the firewall selects for policy application. Layer 4 class-maps sort the traffic based on these criteria listed here. These criteria are specified using the match.where my intention is to let only LAN hosts with IPs in the range 192.168.1.1 to 192.168.1.7 out through the firewall. There may be an easier way of doing this which I'd be pleased to hear about. But, even if there is, I'd also be interested to know what I'm doing wrong in the above.
I have set up a zone-based policy firewall with HA on two 2911 routers as per the Cisco security configuration guide, for an active/passive LAN-LAN cluster. All works as expected, but there is one problem I find: when the control link between the two devices fails, they go into an active/active state as each member assumes it's the last surviving member. The ARP entries for the Virtual IPs on the neighboring devices point to the device that last claimed the active role (usually the standby device). This works in a way, just sessions don't get synched anymore (control link is the same as data link). Now when the link comes back up, the preemtion works and the active, former standby device goes back to standby. But the ARP entries on the neighboring devices still point to the standby device and nothing goes (also sessions established during the active/active state are lost due to resync with the now active member).
This is a single point of failure and what I need is a way to mitigate that. Under:
redundancy application redundancy group 1 control <interface> protocol 1
only one control interface is allowed. Other manufacturers with similar functionality provide for the possibilty of a backup control link, for example the internal LAN interface or a dedicated backup link.
How would I go about that? Maybe use a port-channel for the control/data link (but I'm out of interfaces)?
I have a Cisco 1811 router running the 15.1(3)T IOS. I am having some difficulty with the current zone based firewall and the SSL VPN.
When a user connects, they are put into Virtual-Template 1 which has a zone based assignment of "sslvpn". However the traffic report for the users is listed as being blocked by the zone based firewall in the outbound direction(office out to the wan zone).
The problem I am having is very strange and I have tried to upgrade the IOS on the 1841 to solve the problem but no luck. The issue is when I enable Zone Based firewall security on of the 1841 routers two VPN site-to-site tunnels stops working. If I turn off CEF (no ip cef) then the traffic for both tunnels works. Someone told me that the Zone Based firewall must have a match for the VPN traffic and I created that with ACL 160 and 161 but it did not solve the problem.
Current IOS is below.
Cisco IOS Software, 1841 Software (C1841-ADVSECURITYK9-M), Version 15.0(1)M9, RELEASE SOFTWARE (fc1) Technical Support: [URL] Copyright (c) 1986-2012 by Cisco Systems, Inc. Compiled Tue 11-Sep-12 23:58 by prod_rel_team
I have a Cisco 2901 which terminates a Class C address pool. I have split the Class C address pool into 3 sub-nets and 2 zones and created a non-addressable pool (private pool):
dmz-zone : x.x.x.0 TO x.x.x.127 (x.x.x.0/25) in-zone: x.x.x.128 TO x.x.x.159 (x.x.x.128/27) & x.x.x.160 TO x.x.x.191 (x.x.x.160/27) private-zone: 192.168.x.0 TO 192.168.x.255 (192.168.x.0/24)
I have configured private-zone NAT to use address pool x.x.x.161 TO x.x.x.189 within the in-zone.
dmz-zone - are servers for : DNS, Syslog, SIP & HTTP/HTTPS in-zone - is a SMTP mail server which is behind VPN Gateway/NAT, TomCat (Application Server) and PostgreSQL Server private-zone - is where all standard users are operating from and they can access the SIP & HTTP/HTTPS servers within dmz-zone My problem is that I cannot seem to configure the ZBFW to allow the dmz-zone HTTP/HTTP server to redirect to in-zone TomCat server.
I do not want to make the TomCat server generally visible and am instead using the Apache proxy/ajp13 to connect from dmz-zone server to in-zone server.However I cannot seem to get anything (including icmp) to work from dmz-zone to in-zone.
I have Policy:
POLICY-DMZ-IN (dmz-zone to in-zone) which has: any any udp/tcp inspect any any icmp inspect unmatched traffic DROP/LOG
But I still cannot get anything from dmz-zone to in-zone...Could the POLICY-DMZ-IN be being overridden by other dmz-zone to out-zone policies?
NOTE: I have routing rules for each of various sub-nets and all out-zone to dmz-zone, out-zone to in-zone and private-zone to out-zone, in-zone and dmz-zone routing works ok, so it appears problem is with ZBFW not routing table.
This is problably a stupid question but how do I open a prot on a cisco 1811? I have a cisco 1811 and a computer that has VNC installed on it. I want to be able to access that computer from out side the network using the external ip address and port 5950. People outside the network will be able to open vnc viewer and type in *external ip address*:5950 and it will be directed to the computer with a static internal ip address of 10.11.101.10. What commands do I use to do this?
I have a 2800 router connecting a small office to the Internet. I am using zone-based firewall to provide protection. The small office also needs to connect to another office. The 2800 is at the small office and an ASA at HQ. I successfully established the VPN connection and have allowed Internet access for the small office. The purpose of this post is my zone-base fw policy doesn't appear to be as secure as it could be.
2800 - I have defined two zones (inside and outside). Traffic from the inside to the outside is inspected expect for the traffic to the other office. I allow traffic to the other office to "pass" zbfw. Because the traffic "passes" zbfw, I have to "pass" the same traffic for the outside to in policy. The ASA has "sysopt" to allow VPN traffic to bypass the outside_acl. Do routers and zone-based firewall have a similar feature?
I have a 3845 ISR that I have been managing for a couple years that has a traditional ACL based config. We just purchased a new 3845 for redundancy and it arrived with the zone based config from Cisco. Any opinions on whether I should take the existing router to a zone based config or should I configure the new router with traditional ACL config that I am more comforatable with?
I need to NAT a port range spanning from TCP and UDP 50,000 to 59,999 from inside global address 58.96.x.x on loopback2 to an inside local address of 192.168.5.5.Currently all the existing NAT translations are 1-to-1 that map inside global addresses on a wide span of Loopbacks and a Dialer Interface to inside local addresses on few subnets which are fine.I'm using an 1811 with an ADVIPSERVICESK9-M image, version 12.4(6)TS
Last night I had a crack at setting up PBR on my companies Cisco 1811.Joy, I thought, it's actually working. Alas I was wrong, the addresses were getting translated to our ADSLs external ip address but routed over our EFM.What I want to acheive is to send all HTTP(s) traffic from our workstations over the ADSL (FastEthernet1) whilst all other traffic and VPN goes out over our Bonded ADSL (FastEthernet0). There is also a minor failover in place for traffic routed to the ADSL in the route-map PBR_VLAN1. The servers are on IPs 200, 202, 204 and 240.
Anyway, I have re-written the configuration and xxx'd and x.a/b/c'd all the IP addresses I want to keep secret. Need to make sure that the PBR is correct, and will do what I want it to? I have a very small time-frame to get this correct and I dont want to fudge the bucket so to speak.
I have Site(s) Ani....i=1,..10 sites which communicate with site B to access a website/application. That's simple enough. However, the traffic is http well we primarily don't need https on ipsec tunnel right?. But since attacks related to eavesdropping of traffic come a real reality once it gets terminated by the ipsec device on both side.I have two options either to purchase a third-party ssl certificate to encrypt the traffic between two nodes or use a custom made one.I don't want to use a custom made one because this make the browser prompt an ugly untrusted certificate message; its ugly not from security perspective but for clients inconvenience and assuring users confidence in our systems is a critical issue for us. ?
a) How its possible to remove ugly certifcate message from user screen? Does the company need to register its certificate to some kind of CA body? or what ...
b) Due to some tcp acceleration issues, ssl traffic slows down the traffic between the nodes so we only require the encryption to stand just during the initial handshake when the username and password are being validated ; after that we want to revert back to http?
I am testing out some inspection options on an ASA 5505, and I am running into a situation in which applying a http inspection is dropping all outbound http traffic. I get a "protocol violation" error in the logs.
Here is the setup: I'm not sure why the web traffic is getting dropped.
policy-map type inspect dns preset_dns_map parameters message-length maximum client auto
I have one public IP address but multiple local servers that run on the same port. I cannot change the port the clients use to connect to this server, so I can't do a port map in my NAT router. The solution I had in mind, is to filter on source address. If a client from public IP X.X.X.X connects to port Z, I want it to go to internal server 10.10.10.10 and if a client from public IP Y.Y.Y.Y connects to port Z, I want it to go to internal server 10.20.20.20. Is this possible? I'm using an ASA5510 but I could also switch to a 5505 for this.
I have a customer who wants his new ASA-5520 to load balance out-going traffic between 2 ISPs, fairly normal request. Now here's the twist. He wants to separate traffic based upon the protocol used, http to one ISP, https to the other.
I want to be able to gather some time metrics based on source IP, and destination port. Is it possiable to track how much time a user spends using a service based on it's port number. I have figured out how to capture all the data, and I can then look at timestamps, but I would like a better way if possible. Can this be done at the firewall, or do I need a different appliance?
We are going to "publish"(I don't know if this is correct word to use;)) our mail server on Cisco 2921. As far as I know it can be easily achieved with static NAT. But the thing is, we don't want to publish it on standard 443 port i.e. we want router to listen for https connections on other port than 443, and then redirect this connection to internal server with private ip.
we would like to setup FTP server over CSS where our member sever use non-std-port to open both control/data channel (i.e. 6370 as ctrl and 6369 as data this case.) but seems we only get Passive mode FTP mode work only but not for Active mode FTP case for data channel establishement for server back to client..
# sh ver Version: sg0820501 (08.20.5.01) Flash (Locked): 08.10.1.06 Flash (Operational): 08.20.5.01 Type: PRIMARY Licensed Cmd Set(s): Standard Feature Set
I’m having serious issues getting Tandberg H.323 working behind this router with NAT.
My setup is Cisco 1811 configured with Fas0 to pull DHCP (public address). This router is being used in a mobile medical clinic VAN so the setup needs to be seamless and transparent to the users. The idea with the DHCP is anywhere they go they could pull a DHCP address and then NAT behind that address. The van visits mostly small schools in the Texas Rio Grande Valley providing medical assistance and consulting to the local community. The router has an 8 port built in switch and all ports are sitting in default VLAN 1.
Basic stripped down config, only relevant commands listed…
ip dhcp excluded-address 10.0.0.1 10.0.0.4
ip dhcp pool VANnet network 10.0.0.0 255.255.255.240 default-router 10.0.0.1 dns-server 10.0.0.1(code)
Now initially I can’t even get the call to connect with just using the ports above, which I should. Also knowing there are several issues with H.323 and NAT I went ahead and added all know ports Tandberg says they use…
Basically I created static NAT entries for all the ports and the ranges above. For the ranges I had to add a line for every port.
This didn’t and hasn’t worked yet even with some additional tweaking… Finally the question… am I going about this all wrong? Is there an arrangement of commands that will even work? How can I accomplish the port forwarding setup on a Linksys/Netgear router on a real Cisco router?