BWCTL New Protocol
BWCTL 2.0 has a REST-ish protocol that is used to schedule a network measurement between two servers.
There are 3 entities involved in a BWCTL test: the BWCTL client, the BWCTL A server and the BWCTL B server. The BWCTL client is the entity that requests the two servers perform the test, and retrieves the results from the servers. The BWCTL sender server and BWCTL receiver server are the two servers that perform the test between themselves. The BWCTL servers will each verify that the other's test definition matches their own definition, and that their times are not too far offset. The BWCTL servers will perform the requested test amongst each other, and make the results available for the client.
The BWCTL client retrieves the server status from both BWCTL servers, and makes sure that the server times aren't too far offset, that the test being requested is possible, and what version of the protocol each server speaks.
The BWCTL client requests a test from the BWCTL A server. The server either rejects the test, or accepts the test, and assigns it a time slot. The BWCTL client requests the same test, with the new time slot, from the BWCTL B server. The BWCTL B server either rejects the test, or accepts the test, possibly assigning it a new time slot. If the BWCTL B server rejects the test, the bwctl client cancels the test on the BWCTL A server. If the BWCTL B server changes the time of the test, or changes it's parameters, the BWCTL client submits the changes to the BWCTL A server, who can either accept or reject the changes, or assign it a different time, or other parameters. This back and forth continues until either both servers agree on the time, and parameters, or the test has been deemed impossible given time/parameter constraints.
Once both servers have agreed on the time and parameters, the client updates the test on each server to say that it accepts the test. At this point, each server connects to the other server and verifies that the test they've accepted matches the test that was accepted on the other server, and that their time isn't too far offset from the other server's time. The servers each set their test statuses to accepted, and then check the remote side to make sure that it has accepted the test as well.
After the two servers have both verified and accepted the test parameters, they perform the requested test at the requested time. The client polls the servers until either side finishes. If one side ended in failure, it should cancel the test on the other side, but can let the other side finish up on its own. As each server finishes, the client retrieves the results from that server.
| Endpoint | /bwctl/status |
|---|---|
| GET | Retrieve server status |
The server status consists of an associative array with the following variables:
| Variable | Format | Meaning |
|---|---|---|
| protocol | Floating-Point Number | The version of the protocol that the server understands. Currently, 2.0. |
| time | Millisecond-Grandularity ISO Timestamp | The server's current time |
| ntp_error | Floating-Point Number | The server's maximum NTP error, in seconds |
| available_tools | Array Of Strings | A list of the tools available from the server |
| version | String | The version of BWCTL this server is running |
| Endpoint | /bwctl/tests |
|---|---|
| GET | Retrieve list of tests |
| POST | Request new test |
| Endpoint | /bwctl/tests/<id> |
|---|---|
| GET | Retrieve test description |
| PUT | Update test parameters |
| Endpoint | /bwctl/tests/<id>/cancel |
|---|---|
| POST | Cancel test |
| Endpoint | /bwctl/tests/<id>/accept |
|---|---|
| POST | Accept test |
| Endpoint | /bwctl/tests/<id>/results |
|---|---|
| GET | Retrieve test results |
Tests are defined as associative arrays with the following variables:
| Variable | Format | Meaning |
|---|---|---|
| id | String | The server identifier for the test |
| client_status | String | The status of the test from the client's perspective. "pending" and "accepted" are the two valid options. |
| server_status | String | The status of the test from the server's perspective. "pending", "rejected", "accepted", "finished" and "cancelled" are the valid options. |
| client | Client Object | The parameters about the client making the request |
| sender_endpoint | Endpoint Object | The parameters for the server doing the send side of the test |
| receiver_endpoint | Endpoint Object | The parameters for the server doing the receive side of the test |
| tool | String | The tool type to use for the test |
| tool_parameters | Tool-Specific Parameters Object | The tool parameters to use for the test |
| scheduling_parameters | Scheduling Parameters Object | The scheduling parameters for the test |
Client parameters are defined as associative arrays with the following variables:
| Variable | Format | Meaning |
|---|---|---|
| protocol | Floating-Point Number | The version of the protocol that the client understands. Currently, 2.0 |
| time | Millisecond-Grandularity ISO Timestamp | The client's current time |
| ntp_error | Floating-Point Number | The client's maximum NTP error, in seconds |
Endpoints are the ends of a test. They are defined as associate arrays with the following variables:
| Variable | Format | Meaning |
|---|---|---|
| address | String with IPv4 or IPv6 address | The IP address for the endpoint |
| test_port | Number | The port to use to perform the test |
| bwctl_protocol | Floating-Point Number | The bwctl protocol spoken by this endpoint. There are two special cases: 0.0 means the endpoint cannot be contacted (e.g. there isn't a remote bwctl endpoint), and 1.0 means the endpoint is running bwctl 1.x |
| legacy_client_endpoint | 0 or 1 | Only valid for bwctl 1.x endpoint. If 0, this endpoint is a server endpoint. If 1, this endpoint is a client endpoint. |
| peer_port | Number | The port to use to contact the endpoint |
| base_path | String | Only valid for bwctl 2.x endpoints. The URI path portion of the URL to contact the bwctl server (e.g. '/bwctl') |
| test_id | String | Only valid for bwctl 2.x endpoints. The identifier on this endpoint for the test to be performed |
| posts_endpoint_status | 0 or 1 | Only valid for bwctl 2.x endpoint. If 0, the other endpoint can contact this endpoint to retrieve test/server status information. If 1, the other endpoint must wait for this endpoint to post its test/server status information |
| client_time_offset | Floating-Point Number | The time difference, in seconds, between the endpoint and the client |
Scheduling parameters are used as part of scheduling the test. They are defined as associative arrays with the following variables:
| Variable | Format | Meaning |
|---|---|---|
| priority | Number | Defines the priority of this test. Higher priority tests will preempt lower priority tests. The default priority, if unspecified, is 0. Negative number are valid |
| requested_time | Millisecond-Grandularity ISO Timestamp | The requested start time for a given test |
| latest_acceptable_time | Millisecond-Grandularity ISO Timestamp | The maximum start time that the client will accept |
| accepted_time | Millisecond-Grandularity ISO Timestamp | The actual start time that the server accepted for a given test |
Tool parameters objects define the tool parameters for a given test. They are defined as associate arrays. The acceptable variables are dependent on which tool was selected. However, there are 3 classes of tools, and each tool in a given class shares all accepted variables with the other tools in the class, unless otherwise specified.
| Class | Tools |
|---|---|
| Throughput | iperf, iperf3, nuttcp |
| Latency | ping, owamp |
| Traceroute | traceroute, tracepath, paris-traceroute |
| Variable | Format | Meaning |
|---|---|---|
| duration | Floating-Point Number | The number of seconds that the test will last |
| protocol | String | The protocol to use for the throughput test, all tools currently support "tcp" or "udp" |
| bandwidth | Integer | The bandwidth, in bps, of the test, all tools currently support setting UDP bandwidth. Only iperf3 supports setting TCP bandwidth. |
| parallel_streams | Integer | The number of parallel streams to use for the test |
| report_interval | Floating-Point Number | The number of seconds between when the periodic bandwidth reports |
| window_size | Integer | The TCP window size, in bytes |
| buffer_size | Integer | The length of buffer to read/write, in bytes |
| omit_seconds | Floating-Point Number | Only valid for iperf3. The number of initial seconds to skip when doing bandwidth calculation and reporting |
| tos_bits | Bit Field String | A bit field defining the TOS bits for the test |
| units | String | Defines the bandwidth units for the tool. Meaning is tool-specific |
| output_format | String | Defines the output format for the tool. Meaning is tool-specific |
| receiver_connects | 0 or 1 | Only valid for iperf3 and nuttcp. 0 means that the sending tool connects to the receiving tool (default). 1 means that the receiving tool connects to the sending tool. |
| Variable | Format | Meaning |
|---|---|---|
| maximum_duration | Floating-Point Number | The maximum number of seconds that the test can last |
| protocol | String | The protocol to use for the traceroute test, all tools currently only support "udp" |
| packet_size | Integer | The size, in bytes, of the packets to send, not including the packet headers |
| first_ttl | Integer | Only valid for traceroute and paris-traceoute. The minimum TTL to report |
| last_ttl | Integer | Only valid for traceroute and paris-traceoute. The maximum TTL to report |
| Variable | Format | Meaning |
|---|---|---|
| packet_count | Integer | The number of packets to send in this test |
| inter_packet_time | Floating-Point Number | The time, in seconds, to wait between successive packet sends |
| packet_size | Integer | The size, in bytes, of the packets to send, not including the packet headers |
| packet_ttl | Integer | Only valid for ping. The TTL to use when sending |
| receiver_connects | 0 or 1 | Only valid for owamp. 0 means that the sending tool connects to the receiving tool (default). 1 means that the receiving tool connects to the sending tool. |
Tests results are defined as associative arrays with the following variables:
| Variable | Format | Meaning |
|---|---|---|
| status | String | The test status, one of "accepted", "finished", "failed" or "rejected" |
| results | Results Object | The results of the test itself |
| bwctl_errors | Array of BWCTL Error Objects | The errors that BWCTL encountered while trying to execute this test |
Results objects are defined as associative arrays. Individual tools may include their own tool-specific variables in the array, which may be ignored by the client. The Results object must have the following variable defined:
| Variable | Format | Meaning |
|---|---|---|
| tool_results | String | The combination of stdout and stderr for the given tool that was executed |
BWCTL error objects are defined as associative arrays with the following variables:
| Variable | Format | Meaning |
|---|---|---|
| error_code | Integer | An error code from the big list of errors |
| error_msg | String | A human-readable string description of the error |
In the following example, there are 3 entities: the client, server 1 (192.168.0.2) and server 2 (10.0.0.2).
The client does an HTTP GET request to Server 1 at http://192.168.0.2/bwctl/status.
Response:
{
"version": 2.0.1,
"protocol": 2.0,
"time": "2014-11-18T06:37:03.5678Z",
"ntp_error": 1.2,
"available_tools": [ "iperf", "iperf3", "nuttcp", "traceroute", "tracepath", "paris-traceroute", "owamp", "ping" ]
}
The client verifies that server 1 has the tools it needs, and that server 1's clock isn't too far offset.
The client does an HTTP GET request to Server 2 at http://10.0.0.2/bwctl/status.
Response:
{
"version": 2.0.1,
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.5,
"available_tools": [ "iperf", "iperf3", "nuttcp", "traceroute", "tracepath", "paris-traceroute", "owamp", "ping" ]
}
The client verifies that server 2 has the tools it needs, and that server 2's clock isn't too far offset.
The client does an HTTP POST request to http://192.168.0.2/bwctl/tests.
Request:
{
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requested_time": "2014-11-18T06:37:04.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
The server checks its time, and the client's time, and makes sure they're not too far offset. It then schedules the test, and assigns it an id and a test port number.
Response:
{
"id": "824b8c78-66ae-11e4-9011-001b214e9203",
"client_status": "pending",
"server_status": "pending",
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requested_time": "2014-11-18T06:37:04.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
1415384700,
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
The client checks that the response is acceptable, and then makes a request to server 2, using the 'accepted_time' from server 1 as the 'requested_time' to server 2.
The client does an HTTP POST to http://10.0.0.2/bwctl/tests.
Request:
{
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requesteded_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
The server checks its time, and the client's time, and makes sure they're not too far offset. It then schedules the test, and assigns it an id and a test port number.
Response:
{
"id": "dcf29080-66af-11e4-9c50-001b214e9203",
"client_status": "pending",
"server_status": "pending",
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"test_port": 5044,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "dcf29080-66af-11e4-9c50-001b214e9203",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requesteded_time": "2014-11-18T06:37:05.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
The client checks that the response is acceptable, and then updates the test on server 1 to let it know the "test_id" and "test_port" of the receiver endpoint.
The client does an HTTP PUT to http://192.168.0.2/bwctl/tests/dcf29080-66af-11e4-9c50-001b214e9203.
Request:
{
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"test_port": 5044,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "dcf29080-66af-11e4-9c50-001b214e9203",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requesteded_time": "2014-11-18T06:37:05.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
Response:
{
"id": "824b8c78-66ae-11e4-9011-001b214e9203",
"client_status": "pending",
"server_status": "pending",
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"test_port": 5044,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "dcf29080-66af-11e4-9c50-001b214e9203",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requested_time": "2014-11-18T06:37:04.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
The client verifies that nothing has changed from server 1, and then accepts the test on server 1.
It does an HTTP POST to http://192.168.0.2/bwctl/tests/824b8c78-66ae-11e4-9011-001b214e9203/accept
Request:
{
}
Response:
{
"client_status": "accepted",
}
The client accepts the test on server 2.
It does an HTTP POST to http://10.0.0.2/bwctl/tests/dcf29080-66af-11e4-9c50-001b214e9203/accept
Request:
{
}
Response:
{
"client_status": "accepted",
}
Server 1 does an HTTP GET request to Server 2 at http://10.0.0.2/bwctl/status.
Response:
{
"version": 2.0.1,
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.5,
"available_tools": [ "iperf", "iperf3", "nuttcp", "traceroute", "tracepath", "paris-traceroute", "owamp", "ping" ]
}
Server 1 checks its time, and makes sure that server 2 isn't too far offset.
Server 1 does an HTTP GET request to Server 2 at http://10.0.0.2/bwctl/tests/dcf29080-66af-11e4-9c50-001b214e9203.
Response:
{
"id": "dcf29080-66af-11e4-9c50-001b214e9203",
"client_status": "accepted",
"server_status": "pending",
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"test_port": 5044,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "dcf29080-66af-11e4-9c50-001b214e9203",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requesteded_time": "2014-11-18T06:37:05.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
Server 1 makes sure that the time that server 2 has schedule aligns with its time, and that the test parameters match it's own. It then sets the "server_status" field of its test to "accepted".
Server 2 does an HTTP GET request to Server 1 at http://192.168.0.2/bwctl/status.
Response:
{
"version": 2.0.1,
"protocol": 2.0,
"time": "2014-11-18T06:37:03.5678Z",
"ntp_error": 1.2,
"available_tools": [ "iperf", "iperf3", "nuttcp", "traceroute", "tracepath", "paris-traceroute", "owamp", "ping" ]
}
Server 2 checks its time, and makes sure that server 1 isn't too far offset.
Server 2 does an HTTP GET request to Server 1 at http://192.168.0.2/bwctl/tests/824b8c78-66ae-11e4-9011-001b214e9203.
Response:
{
"id": "824b8c78-66ae-11e4-9011-001b214e9203",
"client_status": "accepted",
"server_status": "pending",
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"test_port": 5044,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "dcf29080-66af-11e4-9c50-001b214e9203",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requested_time": "2014-11-18T06:37:04.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
Server 2 makes sure that the time that server 1 has scheduled aligns with its time, and that the test parameters match it's own. It then sets the "server_status" field of its test to "accepted".
Server 1 does an HTTP GET request to Server 2 at http://10.0.0.2/bwctl/tests/dcf29080-66af-11e4-9c50-001b214e9203.
Response:
{
"id": "dcf29080-66af-11e4-9c50-001b214e9203",
"client_status": "accepted",
"server_status": "accepted",
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"test_port": 5044,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "dcf29080-66af-11e4-9c50-001b214e9203",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requesteded_time": "2014-11-18T06:37:05.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
Server 1 verifies that the "server_status" on Server 2's test is set to 'accepted'. If it is not, server 1 periodically retries until it does.
Server 2 does an HTTP GET request to Server 1 at http://192.168.0.2/bwctl/tests/824b8c78-66ae-11e4-9011-001b214e9203.
Response:
{
"id": "824b8c78-66ae-11e4-9011-001b214e9203",
"client_status": "accepted",
"server_status": "accepted",
"client": {
"protocol": 2.0,
"time": "2014-11-18T06:37:04.12345Z",
"ntp_error": 0.8
},
"sender_endpoint": {
"address": "192.168.0.2",
"test_port": 5012,
"bwctl_protocol": 2.0,
"ntp_error": 1.2,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "824b8c78-66ae-11e4-9011-001b214e9203",
},
"receiver_endpoint": {
"address": "10.0.0.2",
"bwctl_protocol": 2.0,
"test_port": 5044,
"ntp_error": 0.5,
"peer_port": 80,
"base_path": "/bwctl",
"test_id": "dcf29080-66af-11e4-9c50-001b214e9203",
},
"tool": "owamp",
"tool_parameters": {
"packet_count": 10,
"inter_packet_time": 1.0,
"packet_size": 8000,
},
"scheduling_parameters": {
"requested_time": "2014-11-18T06:37:04.12345Z",
"accepted_time": "2014-11-18T06:37:05.12345Z",
"latest_time": "2014-11-18T06:37:10.5678Z"
}
}
Server 2 verifies that the "server_status" on Server 1's test is set to 'accepted'. If it is not, server 2 periodically retries until it does.
The client periodicly retrieves the tests from each server until both are 'accepted'. It then waits for the test to finish. It may periodicly retry to catch failures early.
At the accepted time, Server 2 (the receiver) spawns an OWAMP server that listens on the specified test port (5044). Server 1 (the sender) then spawns an OWAMP client that connects to the OWAMP server. Once the test ends, each side sets its test status to "finished", and records the results for the test.
The client does an HTTP GET request to Server 1 to see if there are results at http://192.168.0.2/bwctl/tests/824b8c78-66ae-11e4-9011-001b214e9203/results.
Response:
{
"status": "finished",
"tool_results": {
"output": "...the output of owping...",
},
}
The client does an HTTP GET request to Server 2 to see if there are results at http://10.0.0.2/bwctl/tests/dcf29080-66af-11e4-9c50-001b214e9203/results.
Response:
{
"status": "finished",
"tool_results": {
"output": "...the output of owampd...",
},
}