Intrusion Detection & Prevention – The Basics

In another article a discussion of ‘odd behaviour’ by computers and networks broke out. I’d said:

@Another Ian:

I’ve not seen that issue, however, I have seen evidence for some wide scale “probing” of computers by what I presume are nefarious sorts. As the spouse is on a Mac, and I’m non-PC of many flavors, we rarely have anything succeed at ‘crawling in’… However, last night (when NOT on any WordPress sites) the spouse had the Mac “lock up” on attempted opening of her (remote) email account. In my experience, that’s what a Mac does when under attack by a very smart attack that is aimed mostly at PCs and just doesn’t “get it” about Macs… (i.e. the Apple defenses stop it from succeeding but both end up in a death match…) Later that night, my Tablet had “glitch” issues (things slowing / jerky / active when doing nothing and load shows) but didn’t succumb. (Android has some exploits, but like the Mac Mach Linux base the Android Linux base is immune to the usual PC / Windoz attacks and most of the few Linux ones…) A reboot of both cleared them and they behaved the rest of the night.

FWIW, I’m likely to be “low response” for the next day or two as I build and deploy a home IDS (Intrusion Detection System) and make a posting on it…

My router blinky lights show no activity out of expected at this point, so IMHO whatever it was has now “moved on” to other easier PC targets… ( I really like blinky lights… they give a ‘no doubt about it’ indication of unwanted / unexpected network traffic… I have a ‘hub’ that I carry with me to most contracts that, put between the wall and my assigned workstation, gives me personal blinky lights… ;-)

Given the recent ‘take down’ of the East Cost by a set of bot driven I[di}OT devices, it’s clear that a pre-survey & PWN was run on them; so I’m not at all surprised that post event it looks like they are doing a survey for the next set of boxes to exploit (the assumption being that many of the first batch will now be identified and locked down / taken offline…) FWIW, it looks like about 9 PM PST was the window, so call it about start of work day in Eastern Europe… Just sayin’… you might want to watch the clock for 8 to 5 Romanian / Kiev / Moscow time…

I ought to have the IDS posting ready by this weekend, and it will include things you can do on your workstation, not just those that need a Linux box, (Host IDS), so I’d suggest installing that kind of thing and seeing what it shows.

For those wanting to “run ahead” and ‘assemble it yourself’:

https://en.wikipedia.org/wiki/Intrusion_detection_system
http://download.cnet.com/s/intrusion-detection/
http://csrc.nist.gov/publications/drafts/800-94-rev1/draft_sp800-94-rev1.pdf
https://suricata-ids.org/download/

The game is afoot and they are probing actively… time for ‘shields up’ and warp engines on line…

H/t to Another Ian for raising the issue.

I’d started down this path a good while back as it was one of the “things you ought to have” in any network, and I didn’t have one. But I was complacent. All my other general protective behaviours had so far kept things clean and safe. (Not the least of which is rarely to never running Microsoft Windows unless absolutely necessary and typically not using it to browse anywhere – the occasional exception of finding and getting needed software. Most virus writers target Microsoft products, partly due to them being a bit easier and mostly due to them being in very large numbers and generally not professionally managed in homes and small businesses.)

In any case, I’ve been relatively lucky (having made a lot of my own luck). But about 2 weeks ago that changed. An “interior router” of mine had been changed. Not much. But it didn’t have the WiFi interface “up” anymore. I looked it over. Changed the config to what it ought to be and rebooted. Then it was fine.

Now being an “interior router” I’d left the default security in place. It was already protected by the AT&T boundary router ( “internet access” router) and in theory nobody could get to it from outside. I have, sometimes, left an open “guest” access up on WiFi for the neighbors should they need it. Had some “war driver” found this from the street? Had the neighbor kid gotten old enough to not just want to access sites without his parents knowing, but also to start hacking routers? Didn’t know, but OK, time to “lock it down”. I turned on the “known machines only” lockout and changed the password to something more industrial strength, then went on my way.

Then the East Coast DDoS attack happened… it used an IOT Internet Of Things (or I[di]OT of Things…) group of drone machines to pester to death various target sites.

Then a couple of nights ago, the spouse and I both had those ‘odd’ behaviours. Her Mac had the spinning meatball of death hang cursor and my Tablet had some quirky not-quite-right pauses and fails to connect to web pages. Both fixed with a reboot…

OK, further down we had

https://chiefio.wordpress.com/2016/10/26/voter-fraud-voting-irregularities/#comment-73972

Larry Ledwick says:
27 October 2016 at 8:38 pm (Edit)
I have also noticed some very slow page loads over the last 24 hours, and semi lockup (ie long delay to do things but eventually succeeding). Things which should have been just a second or less took long enough for me to start grumbling “would you load the friggen page already” sort of muttering.

These are the current security alerts posted by CERT
https://www.us-cert.gov/ncas/alerts

https://chiefio.wordpress.com/2016/10/26/voter-fraud-voting-irregularities/#comment-73980

Power Grab says:
27 October 2016 at 10:08 pm (Edit)
@ Larry Ledwick: I noticed slow page loading right after Obama turned over the Internet to the other side.

I suspect the ‘slow page loads’ is more a matter of ‘slow DNS resolution’ as that can dominate the actual time to load pages (especially those with lots of advertising and links to other pages… each one can take a few seconds for the DNS lookup depending on how your DNS servers are set up and selected).

It can also be the result of the bots (code robots) that are trolling the internet right now looking for things to attack and take over.

I would guess is it less to do with the internet ICAAN handover and more to do with the bots. ICANN just doles out the top level domains, mostly. It’s more a political / negotiation lag thing and less an operational one. The root servers tend to be spread all over the world with distributed operation:

https://en.wikipedia.org/wiki/Root_name_server

OK, going back to that link given by Larry, that’s where all us Geekly types go to find out “what the f… is going on” as we go looking for “CERT Advisories”:

https://www.us-cert.gov/ncas/alerts

Alerts
Alerts provide timely information about current security issues, vulnerabilities, and exploits. Sign up to receive these technical alerts in your inbox or subscribe to our RSS feed.
2016 | 2015 | 2014 | 2013 | 2012 | 2011 | 2010 | 2009 | 2008
View Alerts Feed
TA16-288A : Heightened DDoS Threat Posed by Mirai and Other Botnets
TA16-250A : The Increasing Threat to Network Infrastructure Devices and Recommended Mitigations
TA16-187A : Symantec and Norton Security Products Contain Critical Vulnerabilities
TA16-144A : WPAD Name Collision Vulnerability
TA16-132A : Exploitation of SAP Business Applications
TA16-105A : Apple Ends Support for QuickTime for Windows; New Vulnerabilities Announced
TA16-091A : Ransomware and Recent Variants
TA15-337A : Dorkbot
TA15-314A : Compromised Web Servers and Web Shells – Threat Awareness and Guidance
TA15-286A : Dridex P2P Malware
TA15-240A : Controlling Outbound DNS Access
TA15-213A : Recent Email Phishing Campaigns – Mitigation and Response Recommendations
TA15-195A : Adobe Flash and Microsoft Windows Vulnerabilities
TA15-120A : Securing End-to-End Communications
TA15-119A : Top 30 Targeted High Risk Vulnerabilities
TA15-105A : Simda Botnet
TA15-103A : DNS Zone Transfer AXFR Requests May Leak Domain Information
TA15-098A : AAEH
TA15-051A : Lenovo Superfish Adware Vulnerable to HTTPS Spoofing
TA14-353A : Targeted Destructive Malware

Each of those is a live link at the CERT site. For more info on any one, ‘hit the link’… For now, just notice that Lenovo is called out by name one up from the bottom for a factory installed “Superfish” adware malware that is prone to being taken over. IMHO, just avoid all Lenovo products (really, anything from China if you can). Up at the top, there’s that line about DDoS (Destributed Denial of Service – think being attacked by a million 1st graders all yammering for attention… you get nothing done..) using “Mirai” and botnets (network of roboticly run machines i.e. the stuff in your home used by others to do the attacks). That’s the thing that’s likely causing the slow down in places. Looking inside it:

https://www.us-cert.gov/ncas/alerts/TA16-288A

I’ve bolded a couple of bits. This is a typical alert. Tells you what is wrong, who / where it is coming from if possible, what it does, how to protect against it if known, etc. The date is about right for when the odd things were happening at home.

Alert (TA16-288A)
Heightened DDoS Threat Posed by Mirai and Other Botnets
Original release date: October 14, 2016 | Last revised: October 17, 2016

Systems Affected

Internet of Things (IoT)—an emerging network of devices (e.g., printers, routers, video cameras, smart TVs) that connect to one another via the Internet, often automatically sending and receiving data

Overview

Recently, IoT devices have been used to create large-scale botnets—networks of devices infected with self-propagating malware—that can execute crippling distributed denial-of-service (DDoS) attacks. IoT devices are particularly susceptible to malware, so protecting these devices and connected hardware is critical to protect systems and networks.

Description

On September 20, 2016, Brian Krebs’ security blog (krebsonsecurity.com) was targeted by a massive DDoS attack, one of the largest on record, exceeding 620 gigabits per second (Gbps).[1 (link is external)] An IoT botnet powered by Mirai malware created the DDoS attack. The Mirai malware continuously scans the Internet for vulnerable IoT devices, which are then infected and used in botnet attacks. The Mirai bot uses a short list of 62 common default usernames and passwords to scan for vulnerable devices. Because many IoT devices are unsecured or weakly secured, this short dictionary allows the bot to access hundreds of thousands of devices.[2 (link is external)] The purported Mirai author claimed that over 380,000 IoT devices were enslaved by the Mirai malware in the attack on Krebs’ website.[3 (link is external)]

In late September, a separate Mirai attack on French webhost OVH broke the record for largest recorded DDoS attack. That DDoS was at least 1.1 terabits per second (Tbps), and may have been as large as 1.5 Tbps.[4 (link is external)]
The IoT devices affected in the latest Mirai incidents were primarily home routers, network-enabled cameras, and digital video recorders.[5 (link is external)] Mirai malware source code was published online at the end of September, opening the door to more widespread use of the code to create other DDoS attacks.

In early October, Krebs on Security reported on a separate malware family responsible for other IoT botnet attacks.[6 (link is external)] This other malware, whose source code is not yet public, is named Bashlite. This malware also infects systems through default usernames and passwords. Level 3 Communications, a security firm, indicated that the Bashlite botnet may have about one million enslaved IoT devices.[7 (link is external)]

Impact

With the release of the Mirai source code on the Internet, there are increased risks of more botnets being generated. Both Mirai and Bashlite can exploit the numerous IoT devices that still use default passwords and are easily compromised. Such botnet attacks could severely disrupt an organization’s communications or cause significant financial harm.

Software that is not designed to be secure contains vulnerabilities that can be exploited. Software-connected devices collect data and credentials that could then be sent to an adversary’s collection point in a back-end application.

Solution

Cybersecurity professionals should harden networks against the possibility of a DDoS attack. For more information on DDoS attacks, please refer to US-CERT Security Publication DDoS Quick Guide and the US-CERT Alert on UDP-Based Amplification Attacks.

Mitigation

In order to remove the Mirai malware from an infected IoT device, users and administrators should take the following actions:

Disconnect device from the network.

While disconnected from the network and Internet, perform a reboot. Because Mirai malware exists in dynamic memory, rebooting the device clears the malware [8].

Ensure that the password for accessing the device has been changed from the default password to a strong password. See US-CERT Tip Choosing and Protecting Passwords for more information.

You should reconnect to the network only after rebooting and changing the password. If you reconnect before changing the password, the device could be quickly reinfected with the Mirai malware.

Preventive Steps

In order to prevent a malware infection on an IoT device, users and administrators should take following precautions:
Ensure all default passwords are changed to strong passwords. Default usernames and passwords for most devices can easily be found on the Internet, making devices with default passwords extremely vulnerable.

Update IoT devices with security patches as soon as patches become available.

Disable Universal Plug and Play (UPnP) on routers unless absolutely necessary.[9]

Purchase IoT devices from companies with a reputation for providing secure devices.

Consumers should be aware of the capabilities of the devices and appliances installed in their homes and businesses. If a device comes with a default password or an open Wi-Fi connection, consumers should change the password and only allow it to operate on a home network with a secured Wi-Fi router.

Understand the capabilities of any medical devices intended for at-home use. If the device transmits data or can be operated remotely, it has the potential to be infected.

Monitor Internet Protocol (IP) port 2323/TCP and port 23/TCP for attempts to gain unauthorized control over IoT devices using the network terminal (Telnet) protocol.[10 (link is external)]

Look for suspicious traffic on port 48101. Infected devices often attempt to spread malware by using port 48101 to send results to the threat actor.

Now as you might guess, slugging a Tb/second or so of traffic around, especially if focused on internet service devices, DNS servers, web servers, or just loading up routers with crap, etc. will cause sporadic slowdowns, failure to load, and other quirky behaviour. It can also happen if the botnet is actively probing your gear and trying to take it over.

This is why, when there is an unexplained fault, sloth, lots of ‘blinky lights’ that can’t be explained, etc. I typically down my internet connection (just pull the wire…) and shut down my devices. Then I can do an orderly bring up and asses on them.

Was my interior WiFi router compromised by this code? I don’t think so. Looks more like the neighbor kid learning the limits of hacking… but maybe. In any case, I’d done the right thing in recovering it due to my own habits.

Was the spousal laptop and my Tablet hit by this? Most likely. As a DDos or perhaps as folks pinging the world with Mirai software now that it is public trying to recruit our machines into their own botnet. In either case, my behaviour cleared the machines and after that they worked as expected. (Both powered down for about 20 minutes, then repowered and assessed.)

But the key bit is those last two lines. Where it says things like “Monitor Internet Protocol (IP) port 2323/TCP and port 23/TCP for attempts”… That’s where an Intrusion Detection System comes into play. It’s the thing that does that kind of monitoring. As noted in my first comment on the other page, you can do some ‘light reading’ here:

https://en.wikipedia.org/wiki/Intrusion_detection_system

CNET has soem IDS things you can download or assess on your own:
http://download.cnet.com/s/intrusion-detection/

And of course there’s a formal paper on it all:
http://csrc.nist.gov/publications/drafts/800-94-rev1/draft_sp800-94-rev1.pdf

Finally, this is the IDS I’ll be installing this weekend:
https://suricata-ids.org/download/

This posting is the “background” and not the “how to”. I have to do it first, then I can write the ‘how to’ ;-)

FWIW, this article does a great job in a readable format of going through various IDS / IPS choices, explaining what makes them different, doing a sketch of strengths and weaknesses, and chooses their favorite. That just happens to be Suricata…

https://www.alienvault.com/blogs/security-essentials/open-source-intrusion-detection-tools-a-quick-overview

I doubt I could improve on it much, so just hit the link and read it. Really. Do it now. I’ll wait…

What’s the only reason for not running Snort? If you’re using Suricata instead. Though Suricata’s architecture is different than Snort it behaves the same way as Snort and can use the same signatures. What’s great about Suricata is what else it’s capable of over Snort. It does so much more that it probably deserves a dedicated post of it’s own. Let’s run down a few of them:

Multi-Threaded – Snort runs with a single thread meaning it can only use one CPU(core) at a time. Suricata can run many threads so it can take advantage of all the cpu/cores you have available. There has been much contention on whether this is advantageous, Snort says No and a few benchmarks say Yes.

Built in Hardware Acceleration – Did you know you can use graphic cards to inspect network traffic?

File Extraction – Someone downloading malware? You can capture it right from Suricata and study it.

LuaJIT – It’s a lot of letters yes, but it’s also a scripting engine that can be used with information from the packets inspected by Suricata. This makes complex matching even easier and you can even gain efficiency by combining multiple rules into one script.

Logging more than packets – Suricata can grab and log things like TLS/SSL certs, HTTP requests, DNS requests

So much more…

I’m more familiar with Snort, but this does look superior, so I’ll be getting it set up for myself.

Now maybe you don’t have a dedicated machine to set up to inspect your network traffic, or don’t have a hub handy so that all traffic to your desktop can be duplicated on the interface of your IDS box. You can run Suricata on your desktop directing inspecting the traffic to / from it. This is a little less secure, since the IDS box is more secure if it isn’t actually doing anything else and is so locked down it’s useless as a desktop, but it is still pretty darned good. Especially for a laptop at Starbucks…

The other half of the solution is inspecting inside your computer to make sure things there are not changed. For that, you use a Host IDS.

Host IDS – Host based IDS systems, or HIDS, work by monitoring activity that is occurring internally on a host.
HIDS look for unusual or nefarious activity by examining logs created by the operating system, looking for changes made to key system files, tracking installed software, and sometimes examining the network connections a host makes. The first HIDS systems were rather rudimentary, usually just creating md5 hashes of files on a recurring basis and looking for discrepancies (File Integrity Monitoring). Since then HIDS have grown far more complex and perform a variety of useful security functions. Also if you need to become compliant to one of the many standards (PCI, ISO, etc..) then HIDS is compulsory.

OSSEC

In the realm of full featured Open Source HIDS tools, there is OSSEC and not much else. Go ahead and google away, I’ll wait. The great news is OSSEC is very good at what it does and it is rather extensible. OSSEC will run on almost any major operating system and uses a Client/Server based architecture which is very important in a HIDS system. Since a HIDS could be potentially compromised at the same time the OS is, it’s very important that security and forensic information leave the host and be stored elsewhere as soon as possible to avoid any kind of tampering or obfuscation that would prevent detection.

I’m not that familiar with OSSEC and may try installing it. “In the old days” we ‘rolled our own’ Host IDS. Doing “hash” computation on key files in off hours, comparing file sizes to saved sizes, checking that file system permissions had not changed from the archived set. Now you can get all that pre-made, and more.

While there are lots of other tools and software and very expensive dedicated hardware you can use industrially (I’ve used more of that…) this looks like a good set for the typical home installation. Suricata and OSSEC.

The next posting will be looking at installing and configuring one or both of them to set up a minimal intrusion detection system for home use. I’ll be doing it on a Raspberry Pi, but might give it a try on the laptop just to see how hard / easy it is under Windows.

As a small ‘look ahead’, here’s the basic setup wiki for Suricata:

https://redmine.openinfosecfoundation.org/projects/suricata/wiki/Basic_Setup

It doesn’t look all that hard (at least not to an experienced sysadmin…) I note his bit, though:

sudo cp suricata.yaml /etc/suricata
[…]

Setting variables
Make sure every variable of the vars, address-groups and port-groups in the yaml file is set correctly for your needs. A full explanation is available in the Rule vars section of the yaml. You need to set the ip-address(es) of your local network at HOME_NET. It is recommended to set EXTERNAL_NET to !$HOME_NET. This way, every ip-address but the one set at HOME_NET will be treated as external. It is also possible to set EXTERNAL_NET to ‘any’, only the recommended setting is more precise and lowers the chance that false positives will be generated. HTTP_SERVERS, SMTP_SERVERS, SQL_SERVERS, DNS_SERVERS and TELNET_SERVERS are by default set to HOME_NET. AIM_SERVERS is by default set at ‘any’. These variables have to be set for servers on your network. All settings have to be set to let it have a more accurate effect.

Next, make sure the following ports are set to your needs: HTTP_PORTS, SHELLCODE_PORTS, ORACLE_PORTS and SSH_PORTS.

Finally, set the host-os-policy to your needs. See Host OS Policy in the yaml for a full explanation.

Yet Another Markup Language… Oh Joy, another “language” to learn… It doesn’t look very hard, and is close to many others, so more like a dialect really. One of those things I’ve avoided. Until now. Oh Well.

https://en.wikipedia.org/wiki/YAML

YAML (/ˈjæməl/, rhymes with camel) is a human-readable data serialization language that takes concepts from programming languages such as C, Perl, and Python, and ideas from XML and the data format of electronic mail (RFC 2822). […] Originally YAML was said to mean Yet Another Markup Language, referencing its purpose as a markup language with the yet another construct, but it was then repurposed as YAML Ain’t Markup Language, a recursive acronym, to distinguish its purpose as data-oriented, rather than document markup.

Lucky for us Systems Admin types, we usually don’t need to learn the whole thing, just a few key sentences and how to change the nouns in them… or the occasional verb.

OK, with that, you now got to see what an evening the life of a sys admin looks like when something goes ‘bump in the net’ and their phone rings or the router hangs or “the internet is down” ;-)

After I’ve digested all this (and cleaned up the kitchen and…) I’ll make my first install / config tests. But that will be for another posting.

Subscribe to feed