Home

Awesome

TimeCrypt

Build Status Java Maven Build

Implementation of TimeCrypt, see the Paper for more details. TimeCrypt is an end-to-end encrypted time-series data store where the storage server does not see any data in the clear. The time-series streams are encrypted on the client-side, and only authorized parties can decrypt and verify queries. Features:

Repository Structure

This repository is split into three parts:

Quickstart with Docker

The easiest way to get a TimeCrypt server running in no time is to start it with Docker using docker-compose.

docker-compose up --build

docker-compose builds the project inside a Docker container, creates a Docker network for the server and Cassandra, and starts both containers.

After a successful build, the TimeCrypt server should be running. To test and connect with the server, one needs to interact with the timecrypt-client. The implementation provides some examples on how to use the java driver, a CLI-client, and a dummy data producer (testbed).

The following command runs (Example1):

docker run --network=timecrypt-network eth/timecrypt example1

To see a dummy data producer in action, run the testbed client in Docker with:

docker run --network=timecrypt-network eth/timecrypt producer

For an interactive CLI client session run:

docker run --network=timecrypt-network -it eth/timecrypt client

The client will automatically create a new key store for the cryptographic material of TimeCrypt.

Building TimeCrypt Libraries

To build the project without Docker you will need the following prerequisites on your system:

The project jar libraries can be build with maven.

mvn package

it will resolve all dependencies and build the project. Afterwards the project ccontains the following java libraries

To install the TimeCrypt libraries in the local maven repository run:

mvn install

If your system does not support AES-NI you can disable it during build with the profile switch (-P!aesni-native) e.g.:

mvn package -P \!aesni-native

Run the TimeCrypt Server

The implementation provides two shell scripts to run a server after a successful build Link.

run_timecrypt_in_memory_server.sh

Runs the TimeCrypt server with in-memory storage (i.e., no persistence).

run_timecrypt_cassandra.sh

Runs the TimeCrypt server with Cassandra as a storage-backend. Inspect the scripts to configure the server parameters accordingly.

CLI Client

timecrypt-client/scripts/run_cli_client.sh

Runs the cli client.

Benchmark Client

timecrypt-client/scripts/run_bench_client.sh

Runs the benchmark client, type -h for displaying arguments.

Java Driver

TimeCrypt provides a Java driver to interact with the TimeCrypt server. To use the driver, add the following library to your project.

<dependency>
    <groupId>ch.ethz.dsg.timecrypt</groupId>
    <artifactId>timecrypt-client</artifactId>
    <version>1.0</version>
</dependency>

Example of how to use the library.

timecrypt-examples/scripts/run_example1.sh

Runs the example. Create a TimeCryptClient object to interact with a TimeCrypt server.

TimeCryptProfile profile = new LocalTimeCryptProfile(null, "myUser", "..", SERVER_ADDRESS, SERVER_PORT);
TimeCryptKeystore keystore = LocalTimeCryptKeystore.createLocalKeystore(null, DUMMY_PASSWORD.toCharArray());
TimeCryptClient tcClient = new TimeCryptClient(keystore, profile);

Create a new time-series stream.

long streamID = tcClient.createStream(
                "HeartRate Stream",
                "This stream stores my heart rate",
                TimeUtil.Precision.ONE_SECOND,
                Collections.singletonList(TimeUtil.Precision.TEN_SECONDS),
                DefaultConfigs.getDefaultMetaDataConfig(),
                DefaultConfigs.getDefaultEncryptionScheme(),
                null,
                new Date(startDate)); 

Insert past data, must be inserted in-order:

InsertHandler handler = tcClient.getHandlerForBackupInsert(streamID, new Date(startDate));
handler.writeDataPointToStream(new DataPoint(new Date(timestamp), 100));
handler.terminate();

Insert live data:

InsertHandler liveHandler = tcClient.getHandlerForLiveInsert(streamID);
liveHandler.writeDataPointToStream(new DataPoint(new Date(), 10));
//... periodically insert 
liveHandler.flush();

Query average, variance, standard deviation over the the time range [startDate, endDate).

Interval result = tcClient.performQuery(streamID, startDate, endDate,
                    Arrays.asList(Query.SupportedOperation.AVG, Query.SupportedOperation.VAR, Query.SupportedOperation.STD),
                    false);
System.out.println(String.format("Result [%s, %s] AVG %f, VAR %f, STD %f", x.getFrom(), x.getTo(), x.getValueAt(0), x.getValueAt(1), x.getValueAt(2)));

Disclaimer

This implementation is a research prototype and should not be used in production.