Simmel is designed to be easily adaptable to any contact tracing implementation. This project represents a reference implementation of one possible contact tracing system.
It is based on Singapore's TraceTogether project.
This relies on Circuit Python for NRF boards, such as Circuit Python for Simmel. Ensure you have this installed on your device, then copy code.py over to it. You can use the USB terminal to interact with it or to monitor tracing.
To install, rename bluetrace.py to code.py and drag it on to
the Simmel flash drive.
You can monitor the output by attaching a serial console to the virtual USB serial port.
This program scans for nearby devices, and queries their service for their token. If a token is brand-new, it displays it on the screen. In the TraceTogether system (known as BlueTrace), tokens are valid for fifteen minutes.
After scanning, it begins advertising its own token for a short time. If a device connects, it will allow the remote device to communicate with us, possibly sending us their own token.
This output shows the advertising data, a remote device's connection, as well as what happens when a remote device connects to us. Finally, it also shows that the RTC is functioning normally, and the garbage collector is performing well.
Scanning for BlueTrace
Discovered Addr: <Address 40:f7:31:36:24:a0> Bytes: b'02010206ffff03383663020a041107f918c24c09fef0806a4f9515fcb32ab8' RSSI: -58
| This is a BlueTrace device!
| Random hash ID: 6501944
| Connecting...
| Device reported a connection was opened
| Enumerating GATT characteristics...
| BLE Properties: 0x18
| Bluetrace Data (AES encrypted, 160 bytes): 9d1f382a95391c4eae1796905cc7a36081bdd102b64ec943e51a31230f2c7f76
2dbb7e4a7d4f70db0278e045d362059bee0a13d2baa307d85fc6e5049923afe4
e4f8e9db191802707799189176d2daf9185c16d7d8bfddacc3f2df2a777e1174
3cce8af9c5dbe6e60e121ab41b0fda350829acaa414ba8c576f9cff13c854af7
6fc4c8d628f3412bbd6478b18442bdc8d3f92dbf99ff7f760dd8e7598795342c
Waiting for connection...
| Connected? True Enabled? True Buffer? 0
CN | 25/5/2020 19:52:31 Uptime: 276015.0 Loops? 7048 GC? 26256/29616 Buffer? False Address: <Address e7:a3:3a:9a:48:b2>
o Disconnected from other device
Scanning for BlueTrace
Waiting for connection...
WC | 25/5/2020 19:52:44 Uptime: 276028.0 Loops? 7059 GC? 38496/17376 Buffer? False Address: <Address e7:a3:3a:9a:48:b2>
| Connected? False Enabled? True Buffer? 0
o Disconnected from other device
Scanning for BlueTrace
Waiting for connection...
WC | 25/5/2020 19:52:51 Uptime: 276035.0 Loops? 7064 GC? 7952/47920 Buffer? False Address: <Address e7:a3:3a:9a:48:b2>
Notice how the token is AES encrypted. This could be decrypted using
the aesio module, as long as we had the proper key. Likewise, we
could form our own messages using a known common AES key, which would
prevent us from having to store full encrypted data.
Alternatively, Simmel could store several months' worth of encrypted tokens to transmit, and log encrypted tokens that it sees, all without needing to deal with encryption keys.
This is just a base. Additions would need to be made to create a full contact tracing solution.
Simmel contains an onboard SPI flash device. This can be addressed using
ordinary Circuit Python classes. Raw SPI is ideal for logging this kind
of data, because this is almost entirely a write-only operation. SPI NOR
can be written without erasing, provided it has already been erased. This
means that many outgoing tokens can be loaded into this database, and then
those tokens can be erased to zero once they're used.
Similarly, discovered tokens can be added to a different part of flash, and entire blocks of discovered tokens can be erased at once as soon as they are no longer needed.
This allows for much tighter packing of tokens, and makes efficient use of SPI NOR primitives.
The virtual filesystem is only about 70 kilobytes. This filesystem is stored on internal flash, in order to dedicate the SPI flash chip to contacts. One way to load tokens onto and off of the SPI flash would be to use some serial transfer protocol such as xmodem. Solutions exist, however they need to be ported to Circuit Python.
The Simmel project was started on April 10, 2020 in response to a request for a hardware design proposal by NLNet. The Simmel project name comes from Georg Simmel, an early researcher into sociology and social distance theory.
The Simmel team is funded through the NGI0 PET Fund, a fund established by NLnet with financial support from the European Commission's Next Generation Internet programme, under the aegis of DG Communications Networks, Content and Technology under grant agreement No 825310.