[Snort-users] Snort production setup design

Davison, Charles Robert cdaviso1 at ...17214...
Fri Dec 18 02:28:42 EST 2015


Sorry I did not clarify… when I was referring to VPC I was referencing the AWS VPC (Virtual Private Cloud), and if you apply the router that I was talking about before it sits on the edge of your VPC. If you don’t want to end up using Snort as a HIDS essentially the best thing to do is to buy that router and tap your traffic, inbound/outbound, of your VPC to your desired Snort device. The router to my knowledge supports HSRP so you could build redundancy between the two devices and set up your tap on each router to send the traffic to another VPC or a local Datacenter or anywhere you want. Your SIEM could then read the alerts from whatever locations you set up. And since you have routing redundancy with HSRP if one routing link fails the data would get sent to the other snort instance. Even though you might have separate data centers to help make this redundant you would only have one SIEM reading the same log data so it would put both alerting results into context. There are many ways to perform what you’re talking about but I made a simple diagram below that explains what we are talking about (also attached). Please forgive my lack of artistry/visio :D

[cid:image002.jpg at ...17409...]
If you do have a SIEM you essentially get a copy of your snort logs sent via syslog, there is an option in the snort.conf to add a syslog server, and if this was at a local data center then you would not have to worry about storage space in AWS. And since the cisco virtual routers support enterprise VPNs you could tie them into your network. If you wanted to do HIDS as well this would be fine… you would want to install OSSEC or Samhain on your computer in the AWS VPC, the traffic could then get passed along the same way to your SIEM for interpretation of host activity. Almost everything in the above example could be done via open source except the routing part, which even if you did find an open source router I would recommend going with the Cisco one due to all the feature available to help our with redundancy. Keep in mind that this whole design would rest on the routers not being able to scale elastically. So as long as you are running a small web service you should be fine, but this would be your bottle neck, and you would have to manually add more routers over time to service requests. To my knowledge there is still not official elastic tap capability in AWS for the edge routing piece if there was that would be awesome. Hopefully this answered your questions. If not let me know what I am missing so I can better understand what you’re asking for.

From: sandeep dubey [mailto:sandeep.sanash at ...11827...]
Sent: Thursday, December 17, 2015 9:37 PM
To: Davison, Charles Robert <cdaviso1 at ...17214...>
Cc: Steven Dracker <steven at ...17402...>; Rodgers, Anthony (DTMB) <RodgersA1 at ...17120...>; snort-users at lists.sourceforge.net
Subject: Re: [Snort-users] Snort production setup design

Thanks Charles and all for your opinions and suggestion.

The server i am running is not in VPC mode, it is in Classic mode, this server is very busy and critical where i don't want to install packages like OpenVPN etc., however i can create a new server for this purpose if it works. I want to install minimal package like snort only  which will run in IDS mode and it's log should go to some server where i should be able to analyse those logs and send alerts. I want to avoid space issue on server or i will have to attache EBS volume to store logs locally.

Performance is a big concern in this case because the servers i am running doesn't support load balance nor HA. It is dedicated 3 servers serving RPC calls. If any of the server goes down service gets impacted.

I am looking for only Opensource solutions.

Is it possible to install Snort on these server in NIDS mode and ship the logs to another server like mentioned here - http://sublimerobots.com/2015/12/snort-2-9-8-x-on-ubuntu-part-4/.
In above like it assumes all on single server, where i want snort on my production servers but logs and other tools mentioned should be on another dedicated server.

Please bear with me for basic beginner questions and doubt, I am very new to this (security) field.

Let me know your suggestion, opinions and guidelines.

On Thu, Dec 17, 2015 at 8:46 PM, Davison, Charles Robert <cdaviso1 at ...17214...<mailto:cdaviso1 at ...17214...>> wrote:
If you want to do a true NIDS and not a HIDS on all your AWS boxes there are a couple of things you can do.

1.       Read this article: https://github.com/Security-Onion-Solutions/security-onion/wiki/CloudClient It describes how you can utilize netsniff-ng as a virtual tap. See below for install instructions:

Installing Netsniff-NG: To install netsniff-ng start by the required dependencies:

sudo apt-get install git build-essential ccache flex bison libnl-3-dev libnl-genl-3-dev libnl-route-3-dev libgeoip-dev libnetfilter-conntrack-dev libncurses5-dev liburcu-dev libnacl-dev libpcap-dev zlib1g-dev libcli-dev libnet1-dev

Next, compile and install netsniff-ng.

git clone git://github.com/netsniff-ng/netsniff-ng.git
cd netsniff-ng
sudo make install

2.       The other option would be to buy a cisco virtual router v1000. This router will sit at the edge of your VPC. You can tap traffic from an entire VPC back to a single NIDs box/cluster, probably the easiest option as well. The only thing you wont be able to see is lateral movement within an environment since your taped traffic will only be the external communications. To cover your assets you will need a HIDs installed on your servers for lateral movement. The only downside to this scenario is that the router that sits on the edge does not elastically scale. The last time I checked the v1000’s get 1GB throughput, bidirectional, at the highest licensing tier. You might be able to manually add more v1000’s if you need to accommodate higher traffic to a particular VPC, but you might want to double check with Cisco on that.

3.       If you have any further questions reach out to me and I would be happy to help.

From: sandeep dubey [mailto:sandeep.sanash at ...11827...<mailto:sandeep.sanash at ...11827...>]
Sent: Thursday, December 17, 2015 7:51 AM
To: Steven Dracker <steven at ...17402...<mailto:steven at ...17402...>>
Cc: Rodgers, Anthony (DTMB) <RodgersA1 at ...17120...<mailto:RodgersA1 at ...17120...>>; snort-users at lists.sourceforge.net<mailto:snort-users at lists.sourceforge.net>

Subject: Re: [Snort-users] Snort production setup design

I have installed and configure OSSEC as agent and server for monitoring the system level changes like files, integrity, log monitoring, packages changes, ports changes etc.

Not monitoring any network related stuff though.

On Thu, Dec 17, 2015 at 8:09 PM, Steven Dracker <steven at ...17402...<mailto:steven at ...17402...>> wrote:
I noticed on the GitHub Security Onion Wiki that it requires a Span Port - “For a production deployment, you'll need a tap or SPAN/monitor port. Here are some inexpensive tap/span solutions:”. These solutions look like they are not compatible with AWS.

My findings are that AWS does not support Span, Tap or Mirror on their network layer to get a copy of traffic to inspect which is needed for Network IDS. so I am confused as to how this solution could be deployed for NIDS on AWS. Same thing holds true for Snort.

I have only been successful finding Host Based IDS solutions for AWS which require an Agent on each node. Either they do the IDS analysis on the node itself or do a “soft-tap” on the host’s network adapter (Not at the VPC Perimeter) and pass it to an IDS manager.

How do you do inline HIDS on AWS is my question. I am coming up with a lot of the same questions out there but no answers.


From: sandeep dubey [mailto:sandeep.sanash at ...11827...<mailto:sandeep.sanash at ...11827...>]
Sent: Thursday, December 17, 2015 9:09 AM
To: Rodgers, Anthony (DTMB) <RodgersA1 at ...17120...<mailto:RodgersA1 at ...17120...>>
Cc: snort-users at lists.sourceforge.net<mailto:snort-users at lists.sourceforge.net>
Subject: Re: [Snort-users] Snort production setup design

Thanks Rodgers for reply,

I am running my production environment on public cloud Amazon Web Services (AWS), where i don't have control for installing iso/img etc.

Is SecurityOnion equivalent to OSSIM ?

On Thu, Dec 17, 2015 at 7:03 PM, Rodgers, Anthony (DTMB) <RodgersA1 at ...17120...<mailto:RodgersA1 at ...17120...>> wrote:
Can’t recommend SecurityOnion highly enough.

Anthony Rodgers
Security Analyst
Michigan Security Operations Center (MiSOC)
DTMB, Michigan Cyber Security

From: sandeep dubey [mailto:sandeep.sanash at ...11827...<mailto:sandeep.sanash at ...11827...>]
Sent: Thursday, December 17, 2015 04:53
To: snort-users at lists.sourceforge.net<mailto:snort-users at lists.sourceforge.net>
Subject: [Snort-users] Snort production setup design


Is it possible to install snort in IDS mode on multiple servers (AWS EC2 instances ) and have a central server where analysis can be done through gui and also alerts/notification can be managed like OSSEC ?

If yes, what is the tools to use and how to move ahead?




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.snort.org/pipermail/snort-users/attachments/20151218/44acad5e/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.jpg
Type: image/jpeg
Size: 26323 bytes
Desc: image002.jpg
URL: <https://lists.snort.org/pipermail/snort-users/attachments/20151218/44acad5e/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: snortaws.JPG
Type: image/jpeg
Size: 58294 bytes
Desc: snortaws.JPG
URL: <https://lists.snort.org/pipermail/snort-users/attachments/20151218/44acad5e/attachment.jpe>

More information about the Snort-users mailing list