ip-reputation-monitoring

Build Status Coverage Status

Summary

This tool aims to monitor the reputation of a network in order to easily detect blacklisted IPs or IPs sending suspicious e-mails. Every day an e-mail is sent telling you the 10th IPs having the worst reputation among the network.

The monitoring is based over both RBL and FBL as described below.

An API is provided in order to query the reputation of a single IP using RBL and FBL aggregation. The API also offers an endpoint that checks DNS black list.

RBL

A Realtime Blackhole List is composed of IP addresses considered as spammers. Several organizations are maintaining such list and this tool supports some of them: Microsoft SNDS, BlockList, CleanTalk, StopForumSpam.

Generally, those lists are released under CSV format. Unfortunately, CleanTalk does only release this list in HTML, so a HTML to CSV parser has been written to do so.

FBL

A feedback loop is a realtime feedback about mails sent from observed network. Those mails can be sent by e-mail hosting providers or by foundations fighting against spams in order to give a feedback about suspicious mails issued from the network. The tool supports Abuse Reporting Format (ARF) which is a standard to report suspicious mails used by AOL and SignalSpam feebacks.

SpamCop feebacks are also supported by the tool, even if their mails do not follows ARF guidelines.

DNS BL

The DNS-based blackhole list is a subset of RBL and, instead of being a CSV file containing an IP list, is a software mechanism allowing anyone to query the database using the DNS protocol whether an IP is blacklist.

Supported DNS BL are the following:

Spamhaus BL

Spamhaus is an organization tracking spam over Internet. Their website provides an easy way to keep an eye over active issues on a network giving the domain responsible of an IP set.

This tool provides a script able to parse the issues raised by Spamhaus to track network reputation.

Installation

Requirements

To setup this little tool, you'll need:

*TLS support can be disabled by editing settings/config.py.

Step by step

When all of these requirements are met, you can install the tool:

  1. Download the zipfile or checkout the sources.
  2. Install python dependencies (apt-get install python-dev python-pip)
  3. Run make install-deps.

Configuration

General settings

For security purpose, most of setting values must be defined as VARENV. So, to configure the tool, you just have to export following environment variables:

Tagging incoming e-mails

As you can see, a required varenv is called FBL_PARTNER_HEADER. Since everybody is able to send an email pretending being Microsoft SNDS, NSA or whatever, it's important to check the identity of the mail send with a well-kept secret. The easier way to do so is to assign a single e-mail address per organization you keep secret.

Using the MX features, the mail must be validated and tagged with the previous FBL_PARTNER_HEADER header. We recommend to keep this header name secret too. The value of this header is the name of the FBL. (currently supported values: AOL, SignalSpam, SpamCop)

An example should help you to understand:

OVH wants to received spam report from AOL.

OVH creates a new e-mail address for AOL: fbl-55eb4d8efa1@ovh.net.
AOL uses it to send all spam report.
OVH adds a new rule attached to this address: when receiving a new e-mail
from  <scomp@aol.net>, add header "X-PARTNER-HEADER: AOL" and forward this
e-mail to the FBL_USER varenv mailbox.

The monitoring tool is polling MX the FBL_USER varenv mailbox on EMAIL_HOST
server and its "FBL_PARTNER_HEADER" has been set to "X-PARTNER-HEADER`.

Network IPs

At last, you need add your network addresses CIDRs (only one per line) in the config/ips.list file.

Implementing its own RBL storage class

As you can see in the config.py, there is a way to customize the tool by providing its own implementation to store RBL parsed documents. By default, a basic implementation is provided and store everything on the filesystem, using the property RBL_STORAGE_CONTEXT to determine the root path to use.

You can code your own RBL storage service by implementing adapters.service.storage.StorageServiceBase and then, tell the tool to use this implementation by editing the property CUSTOM_IMPLEMENTATIONS.

Implementing new parsers

Implementing new parsers is painless and you'll only have to implement an interface. If your parser is valid enough, it should be automatically registered and enabled as long as daemons are restarted.

1. Implementing a new CSV parser

For a brand new CSV parser, you'll need to implement parser.csv.csvparser.CSVParser. Here are the explanations of methods that must be implemented:

# -*- encoding: utf-8 -*-

from datetime import datetime
from parser.csv.csvparser import CSVParser

class MyNewParser(CSVParser):

    def __init__(self, path):
        # Consider the delimiter is a comma.
        CSVParser.__init__(self, path, ',')

    def compute_weight(self, data):
        # The weight of this entry. All the weights are then summed to rank IPs.
        # Notice data is an array containing the splitted line.
        return 1

    def get_date(self, data):
        # Date to use for this entry
        return datetime.now()

    def get_source(self, data):
        # Source name
        return 'My new parser'

    def get_ip(self, data):
        # IP the entry is talking about
        return data[0]

    def get_description():
        # Mandatory method to be automatically registered !
        return {
            'name': 'My new parser',
            'shortened': 'MNP'  # Shortened name, optionnal
        }

    get_description # staticmethod(get_description)

2. Implementing a new mail reader

Unlike CSV parsers, mail reader are not automatically registered. That's why if you only want to add a new provider using ARF, you'll have to edit parsing.mails.mailfactory.MailReaderFactory and add the new source name. (This code can be greatly improved, your PR are welcomed !)

For a brand new mail reader, you'll need to edit this file too to add your source. And, you'll have to implement an abstract class to be able to read mails. Here is the default format:

# -*- encoding: utf-8 -*-

from datetime import datetime
from parsing.mails.mailreader import AbstractMailReader

class MyNewReader(MailReader):

    def __init__(self, raw):
        AbstractMailReader.__init__(self)
        # Raw is the received e-mail.
        self._data # raw

    def compute_weight(self):
        # The weight of this entry. All the weights are then summed to rank IPs.
        return 1

    def get_date(self):
        # Date to use for this entry.
        return datetime.now()

    def get_source(self):
        # Source name
        return 'My new mail reader'

    def get_ip(self):
        # IP the mail is talking about
        return "1.2.3.4"

Extending DNS BL support

If you want to add new DNS BL that are not supported by default, you just have to edit the file config/dnsbl.py and add a new dictionnary providing needed information about the DNS BL.

Note that the shortened name is mandatory.

Running

You can now insert theses entries in your favorite scheduler:

API

Once everything is running, you can start using the API. By default, it's listening to the port 5000. Here are the few available endpoints:

Enjoy