Can anyone suggest ...
How to detect or prevent malware in encrypted traffic without depending on a security tool
I am aware that Cisco comes with Encrypted Traffic Analytics (ETA), which monitors network packet metadata to detect malicious traffic even if its encrypted , but i would like to know any other suggestions for detection and prevention control
Thanks
Solved! Go to Solution.
In that case you will always need the ability to decrypt the incoming transmission, and you may have re-encrypt it again by policy, before it is forwarded on to the final destination. Normally, one would have an assured solution, as it normally has to hold a copy of the private key for decryption purpose. Normally the solution has normally a proxy or an SSL/TLS forwarding capability with layer 4 to layer 7 inspection capabilities.
Regards
Caute_cautim
WAF (Web Application Firewalls) may help to address this issue
These firewalls are specific enough that they know the way the application should be behaving
and can detect even the smallest unusual activity and bring it to a stop. In addition, WAFs
can also provide protection against such network-based attacks as DoS or DDoS attacks.
Reverse and Forward Proxy would help.
any comments please??
Yes, you could do it that way - there also other means, via having a front web application proxy or Mobile aware proxy, with one way key decryption or as you state forward or reverse proxy as well. Also there is appears to be another method - using a cloud based web application API, which some vendors provide as a stop gap, to keep PCI DSS issues arising i.e. TLS V1.0 issues and preventing access and related vulnerabilities.
This is more like a stop gap approach, rather than a permanent, but often see it taken up as a solution, which appears to become the norm,
Regards
Caute_cautim
You could also choose to decrypt traffic at edge firewall, or at the IPS/web proxy in the middle.
Do take consideration on the additional resource overhead it will put on the device so it wont affect its original function.
With the majority of web and Internet traffic encrypted, you are right to be concerned about it being an avenue to malware.
Most web proxy filters (Bluecoat, Zscaler) and firewalls will also filter web pages that are from known bad sites without decryption. Unless you are willing/able to decrypt the communications (e.g using man-in-the-middle techniques) you are pretty much limited to site-level reputation filtering, (e.g. known bad site, young DNS registration, caught hosting malware, etc). With decryption, it becomes possible to delve deeper, such as allowing chat, but to denying file transfer; and also to AV scan individual files. For example, Facebook is "social networking", but with decryption, you get the ability to allow chat while blocking file transfer.
Also useful is to watch for hosts going to known malware"phone home" sites. This gives an indication of which hosts may already be infected and need remediation.
For now, I have only seen Cisco ETA really working when it comes to encrypted traffic analysis. Although some other vendors claimed the ability to detect 0-day threads (e.g. Darktrace or Greycortex) from our testing the best results came from Cisco Stealthwatch with ETA.
Some other options tough are able to do partial job in malware protection as well. But the majority of the function is based on reputation database for destination IPs/domains... (what is unable to discover malware communicating to twitter or instagram for example).
So aside the Cisco ETA, you basically have another two good options:
1. enhance end-point protection, where the communication is initiated and the payload is processed unencrypted (Cisco AMP for endpoints does great job as it is tracking all operations and communication of the endpoint and the infection can be even discovered afterwards and you posses great data for retrospective analysis).
2. implement decryption at proxy in order to inspect payload.
solution at the end point / proxy seems a good choice but, decrypting the traffic has a an impact in terms of time, performance and cost and in some areas is simply not possible because the necessary cryptographic keys aren't available.
The aspect of this approach, however, is that it may infringe the privacy policy.
As others here have mentioned decryption at the edge device is a great way to accomplish this. If you go this route keep the following things in mind:
* there is an impact to the througput/speed (although we have found it to be unnoticeable with the properly sized hardware)
* You will want to take into consideration things that you should *not* decrypt (HIPPA, etc)
* Thick client apps that use certificate pinning / hard coded certs will not play nice and will end up requiring exceptions