Score Engine 2

This is the software used to perform scoring for UBNetDef's Lockdown competition.

Installation

Score Engine 2 can be "installed" in 3 ways:

1. Pulling the docker.pkg.github.com/ubnetdef/scoreengine2/scoreengine2 Docker image (requires Docker Engine) docker pull docker.pkg.github.com/ubnetdef/scoreengine2/scoreengine2
2. Building from the Dockerfile in the repository (requires Docker Engine)
   • docker build -t scoreengine2 .
3. Installing the repository as a Python package
   • Clone the repository and change into that directory
   • Optionally activate a virtual environment
   • pip install .

In either case, you will need to prepare a config file.

Configuration

Score Engine 2 expects a configuration file named config.yml. An example, named config-sample.yml, is provided.

The configuration file consists of a number of mappings (i.e. key: value pairs). Each key is in described in a sub-section having the same name below, and may have a value that is another mapping or a sequence (i.e. list of values).

bank

This is a mapping that configures the Score Engine to use the Bank API

• enabled (boolean): whether or not the Bank API should be used
• url (string): the full URL (including the protocol and the path) of the Bank API's endpoint for transferring money to the earning team
• username (string): the username of the Score Engine's Bank API user
• password (string): the password of the Score Engine's Bank API user

Example:

bank:
  enabled: yes
  url: https://bank.example.com/internalGiveMoney
  username: scoreengine
  password: secret

celery

This is a mapping that allows configuring Celery.

• backend (string): see Celery docs
• broker (string): see Celery docs
• worker (mapping): see Celery docs

Example: Use a MySQL database as the results backend, RabbitMQ as a broker, and workers that concurrently process up to 20 tasks, with log level INFO and tracebacks.

celery:
  backend: "mysql+pymysql://username:[email protected]:3306/dbname"
  broker: "amqp://username:[email protected]:5672//"
  worker:
    concurrency: 20
    loglevel: INFO
    traceback: yes

checks

This is a mapping that allows overriding default parameters of groups of checks.

Example: change the ICMP timeout to 5 seconds and the FTP directory to /opt/

database

This is a mapping that will be used by the SQLAlchemy ORM to establish connections to the database engine.

The simplest mapping will specify the database URL (url).

Example:

database:
  url: "mysql+pymysql://username:[email protected]:3306/scoreengine"

See also:

• SQLAlchemy database URL format
• SQLAlchemy create_engine parameters

logging

This is a mapping that configures logging.

• format (string): see LogRecord attributes
• level (string): see Logging Levels

Example:

logging:
  format: '%(asctime)s [%(levelname)s] %(name)s: %(message)s'
  level: INFO

round

This is a mapping that determines how the Score Engine treats each round.

• duration (integer): once we start a round, we wait this many seconds before starting the next round; this does not consider how long it takes for a round to complete
• jitter (integer): each round can start up to this many seconds before or after the expected duration (determined at random); can be used to increase uncertainty about when a round will start

Example: Wait 50 to 70 seconds between the start of each round

round:
  duration: 60
  jitter: 10

services

This is a sequence of mappings that configures the services that are being scored. Note that this is only considered when doing the database setup.

• name (string): Custom name used to identify the service (usually unique).
• check (string): Module and function names, joined by . (full stop or period), in the form <module_name>.<function_name>, e.g. database.check_query_mysql or ftp.check_upload_download:
  • Module name is the file name in ./scoreengine/checks without the .py extension
  • Function name is from within the module (some modules have mutiple functions)
• data (sequence of mappings):
  • key (string):
    • Check-specific identifier for an item of data that might be used in checks
    • If multiple mappings are provided with the same key, one of them will be randomly selected for use
    • USERPASS is "special", in that it will be broken into two mappings (USER and PASS) before being sent to the checks; the value is split at ||
  • value (string):
    • Value corresponding to the key specified by key
    • {team} will be replaced with the team number
  • editable (boolean; default: false): whether blue teams can edit this mapping in the Inject Engine
  • hidden (boolean; default: false): whether this mapping should be hidden in the Inject Engine

Example:

• Check if dns.example.com contains an A record for ad.example.com that maps to 192.168.<team_number>.14
• Check if we can SSH into 192.168.<team_number>.22 with username demo and password changeme (allowing the blue team to change the credentials)

services:
  - name: DNS
    check: dns.check_dns
    data:
      - key: HOST
        value: "dns.example.com"
      - key: LOOKUP
        value: "ad.example.com"
      - key: TYPE
        value: A
      - key: EXPECTED
        value: "192.168.{team}.14"
  - name: Client X SSH
    check: ssh.check_connect
    data:
      - key: HOST
        value: "192.168.{team}.22"
      - key: USERPASS
        value: "demo||changeme"
        editable: yes

teams

This is a mapping that configures the teams to be checked. Note that this is only considered when doing the database setup.

• minimum (integer): the smallest team number to be configured (inclusive)
• maximum (integer): the largest team number to be configured (inclusive)
• check_teams (sequence of integers): the teams that should be marked as check teams (and hence are hidden from the Inject Engine's public score board)

Example: Configure 15 teams ({1, 2, 3, ..., 15}), with 13, 14 and 15 as check teams

teams:
  minimum: 1
  maximum: 15
  check_teams: [13, 14, 15]

trafficgen

This is a mapping that is used to simulate additional (unscored) network traffic that can operate on a different schedule than the scored network traffic. Checks will be performed in the same way as scored checks, but without storing the results and calling the Bank API.

• number (integer): the maximum number of checks scheduled per traffic round
• sleep (integer): the amount of time to wait between traffic generation rounds

Example: Generate up to 20 random checks every 15 seconds

trafficgen:
  number: 20
  sleep: 15

Usage

TODO: This section is in need of improvement (more details and examples). However, the CLI provides documentation about it's commands via --help.

1. Create the configuration file config.yml
   • If using a container, mount the file to /opt/scoreengine2/config.yml
2. Run scoreengine2 as needed (see scoreengine2 --help)