888888888888 88888888ba ,ad8888ba, 88888888ba
,88 ,d ,d 88 "8b d8"' `"8b 88 "8b
,88" 88 88 88 ,8P d8' 88 ,8P
,88" ,adPPYba, MM88MMM MM88MMM ,adPPYYba, 88aaaaaa8P' 88 88aaaaaa8P'
,88" a8P_____88 88 88 "" `Y8 88""""""8b, 88 88888 88""""""'
,88" 8PP""""""" 88 88 ,adPPPPP88 88 `8b Y8, 88 88
88" "8b, ,aa 88, 88, 88, ,88 88 a8P Y8a. .a88 88
888888888888 `"Ybbd8"' "Y888 "Y888 `"8bbdP"Y8 88888888P" `"Y88888P" 88
A valid ExaBGP configuration must be provided that includes a process for ZettaBGP with the json encoder.
You can use this reference as a sample.
neighbor 172.17.179.104 {
router-id 172.17.179.103;
local-address 172.17.179.103;
local-as 1;
peer-as 1;
hold-time 180;
api {
processes [zettabgp];
receive {
parsed;
update;
}
}
}
process zettabgp {
run python3 /home/imprj/zettabgp/src/main.py;
encoder json;
}
For the WebApp you can provide a local MRT Library.
This library contains several MRT Scenarios which define a predefined scenario for simulating MRT files with specific configurations.
This library is managed via scenario.json files in the mrt_library directory.
The syntax for the scenario.json files is the following.
{
"name": "Test Scenario 1",
"description": "This is a test scenario.",
"no_rabbitmq_direct": false,
"rabbitmq_grouped": 5,
"no_mongodb_log": true,
"no_mongodb_state": true,
"no_mongodb_statistics": true,
"clear_mongodb": false,
"playback_speed": null,
"mrt_files": [
"20241005_0130_1728091800_bgp_lw_ixp_decix_update.bz2"
]
}A sample setup for MongoDB and RabbitMQ is provided in the db directory.
To run the docker compose setup run the following command in that directory.\
sudo docker compose up -d
The data will be stored in the data directory.
The following environment variables can be used to specify connection parameters.
RABBIT_MQ_HOST
MONGO_DB_HOST
MONGO_DB_PORT
For internal configurations there are these variables too.
But only overwrite them when you are debugging the respective parts locally.
ZETTABGP_WEBAPP_UI_PATH
ZETTABGP_WEBAPP_APP
ZETTABGP_WEBAPP_MRT_LIBRARY_PATH
ZettaBGP provides a CLI interface with some commands for testbed simulations as well for production use.
The exabgp subcommand is used for processing ExaBGP Messages.
The process can be started from within ExaBGP.
The Messages will be received using the stdin pipe.
Options:
-d, --no-rabbitmq-direct
-g, --rabbitmq-grouped INTEGER Queue group interval in minutes. [default: (5)]
-l, --no-mongodb-log
-s, --no-mongodb-state
-t, --no-mongodb-statistics
-c, --clear-mongodb
Besides the direct RabbitMQ queue (Exchange: zettabgp with Routing Key: direct), a grouped queue can be activated.
The grouped queue (Exchange: zettabgp with Routing Key: grouped) will be filled white grouped route updates when enabled with the -g option.
The group interval defaults to 5 minutes.
Alternative intervals can be set as an argument to option -g.
A 10 minute interval can be set as following.
zettabgp exabgp -g 10
When no -g option is present, no grouped updates will appear at all.
The mrt-simulation subcommand is used for processing mrt files.
It is mandatory to provide a valid path to at least one mrt file.
mrt-simulation also supports the handling of multiple mrt files.
But keep in mind to provide sequentially sorted mrt files based on the timeframe.
Otherwise the grouping feature will not work properly!
Arguments:
MRT_FILES...
Options:
-d, --no-rabbitmq-direct
-g, --rabbitmq-grouped INTEGER Queue group interval in minutes. [default: (5)]
-l, --no-mongodb-log
-s, --no-mongodb-state
-t, --no-mongodb-statistics
-c, --clear-mongodb
-p, --playback-speed INTEGER Playback speed in multiples of real time. [default: (1)]
-o, --playback-interval INTEGER Playback interval in minutes. [default: (5)]
Without specifying a playback speed, mrt-simulation will replay all route updates at once.
When defining playback speed, the replay of the updates will be done in multiples of real time.
For a real time playback, you can use option -p without an argument.
zettabgp mrt-simulation <mrt-file> -p
For a playback speed that is twice as fast as real time, the option -p can be used with argument 2 (2x speed of real time).
zettabgp mrt-simulation <mrt-file> -p 2
For debugging the timebase group update queue, it is very useful to playback all update messages that occur within an interval of for example 5 minutes.
When you specify option -o you can set a playback interval in minutes that defaults to 5 minutes.
Between the intervals you have to press enter to continue with the replay of the next interval.
Of course you can combine this option with the playback speed option.
An 5 minute interval playback looks like that.
zettabgp mrt-simulation <mrt-file> -o
A 10 minute interval playback looks like that.
zettabgp mrt-simulation <mrt-file> -o 10
Please keep in mind that most of the mrt files only contain a timeslot of 15 minutes.
This command launches the debugging WebApp with MRT Library, Anomaly Explorer and the RabbitMQ interface.
The WebApp will be exposed on port 8000 by default.
You can access the ui using this link: http://127.0.0.1:8000/
For the management of the MRT Library refer to this section above. Local MRT Library
This command lets you import the contents of rib files, similar to the mrt-simulation.
A path to a rib file is mandatory for this command.
This command uses a part of the options as mrt-simulation does.
Arguments:
RIB_FILE
Options:
-d, --no-rabbitmq-direct
-g, --rabbitmq-grouped INTEGER Queue group interval in minutes. [default: (5)]
-l, --no-mongodb-log
-s, --no-mongodb-state
-t, --no-mongodb-statistics
-c, --clear-mongodb
This command lets you load already saved BGP messages from the database and replays them.
You can use all options as in the mrt-simulation command.
Additionally you can give a timeframe for the messages you want to replay.
Options:
-d, --no-rabbitmq-direct
-g, --rabbitmq-grouped INTEGER Queue group interval in minutes. [default: (5)]
-l, --no-mongodb-log
-s, --no-mongodb-state
-t, --no-mongodb-statistics
-c, --clear-mongodb
-p, --playback-speed INTEGER Playback speed in multiples of real time. [default: (1)]
-o, --playback-interval INTEGER Playback interval in minutes. [default: (5)]
-b, --start-timestamp FLOAT Timestamp for the starttime of the replay
-e, --end-timestamp FLOAT Timestamp for the endtime of the replay
-r, --start-time STRING Time for the starttime of the replay, format (T is a set character): YYYY-MM-DDThh:mm:ss
-f, --end-time STRING Time for the endtime of the replay, format (T is a set character): YYYY-MM-DDThh:mm:ss
When using the -b option you must use the -e option aswell.
For -r and -f the same logic applies.
When no timeframe is provided or only one of the two necessary options is set, the whole database will be loaded and used for the replay.
Some sample json messages for debugging purposes from ExaBGP can be found in the samples directory.
When you want to access the MongoDB and RabbitMQ instances on the testbed, you can use a remote port forward with ssh to forward the application ports to your local machine.
ssh -L 15672:127.0.0.1:15672 -L 5672:127.0.0.1:5672 -L 27017:127.0.0.1:27017 node103
class ChangeType(Enum):
ANNOUNCE = 1,
WITHDRAW = 2,
class NLRI(BaseModel):
prefix: str
length: int
class OriginType(Enum):
IGP = 1
EGP = 2
INCOMPLETE = 3
class AsPathType(Enum):
AS_SET = 1
AS_SEQUENCE = 2
AS_CONFED_SET = 3
AS_CONFED_SEQUENCE = 4
class AsPath(BaseModel):
type: AsPathType
value: list[int]
class Aggregator(BaseModel):
router_id: str
router_as: int
class PathAttributes(BaseModel):
origin: Optional[OriginType] = None
as_path: Optional[list[AsPath]] = None
next_hop: Optional[list[str]] = None
multi_exit_disc: Optional[int] = None
local_pref: Optional[int] = None
atomic_aggregate: Optional[bool] = None
aggregator: Optional[Aggregator] = None
community: Optional[list[list[int]]] = None
large_community: Optional[list[list[int]]] = None
extended_community: Optional[list[str]] = None
orginator_id: Optional[str] = None
cluster_list: Optional[list[str]] = None
class RouteUpdate(BaseModel):
timestamp: datetime = datetime.now()
peer_ip: str
local_ip: str
peer_as: int
local_as: int
path_attributes: PathAttributes
change_type: ChangeType = None
nlri: NLRI = None