Scientists collect massive amounts of data. In fact, the biggest challenge of science is sifting through all the data to come to correct and fascinating conclusions. This is where you come in: If you want to help with the march of science, you could give your life to NASA or a large university, or you could just contribute some cycles from your home computer. One of the easiest ways to contribute to science is to let your computer crunch the numbers for ongoing scientific experiments.

It all started with the search for extraterrestrial intelligence (SETI). SETI scientists built a radio telescope to scan the skies for radio signals that could come from civilizations like ours. The big problem quickly became obvious; analyzing the data takes immense amounts of computing power. To address this problem, the team adapted the software they use to analyze the signal to run on a PC. Then they came up with a brilliant idea: make small work packages that anyone can compute in any order. Users can download the packages and analyze the data off-line. This client was then shared on their website, the best part: they designed it so users could use the software as a screen saver.

The popularity of the project exceeded their expectations, and the team quickly realized that the system could work for other projects also. However, other projects such as medical research and astronomy needed other mathematics. The solution was Berkeley Open Infrastructure for Network Computing, BOINC [1].

BOINC is a framework for crowd-sourced scientific analysis projects. Volunteers can download software and then choose which project they want to support. Institutions can even write their own code and add it to the list of supported projects.

The BOINC Client

Installing the BOINC client on most distributions requires only your favorite package manager. You'll need to install two packages. The first package is the client that does all the work. The other package is the GUI where you define which projects you wish to support. Without the GUI, you will have to configure your projects at the command line.

If you prefer to compile the BOINC client yourself, all the source code is available on GitHub [2]. Cloning the files to your machine is one way to get started; the git repository has the source for both the client, the manager, and the API. Once the client is installed, you can start crunching data, host a project, or even create your own project for others to help you. If you create your own project, you can save work by creating a VirtualBox image to run it in.

To install BOINC in Ubuntu, use the apt command.

sudo apt install boinc

The BOINC package is a meta package that installs the client and the GUI manager. It does not install boinctui, which is curses based and can be useful for servers where only ssh connections available. Other packages you need to consider are the ones that let BOINC access the GPU. The packages will vary depending on your hardware; for Nvidia, you can use CUDA. BOINC has its own package for utilizing these libraries; boinc-client-nvidia-cuda is a meta package that takes care of installing everything necessary.

If you wish to use BOINC to start your own project, it is a good idea to use a virtual machine, because you'll need many special packages, and you might not want to make all these changes to your production system.

You can run BOINC with the command-line tool called boinccmd, which is useful if you want to create your own script. If, however, you just want to add a project or two, use boincmgr. The boincmgr manager tool is a GUI that lets you choose to start, suspend, and stop projects (Figure 1).

Figure 1: Choose your projects using the graphical BOINC manager.

BOINC Projects

Once BOINC is up and running, you can choose to support many projects with one client. SETI is fascinating, but what about helping with other research? BOINC supports a number of medical research projects. Denis@home simulates electrical heart activity. Rosetta@home simulates protein – the results are useful for research on HIV, malaria, cancer, and dementia. FERMI lab has also weighed in to get help analyzing data from the Large Hadron Collider. Other physics projects are also available. One noteable example is Einstein@home, which analyses LIGO data to search for gravitational waves. If you want to explore our galaxy, join Milkyway@home. To go further, join Universe@home to help create a simulation of the entire universe.

Several projects even let you collect data yourself, which requires the purchase of hardware. The hardware is usually cheap though. For instance, the radioactive@home project sells their sensor for E25 with a rebate for large orders.

The Folding@home project comes as two different files: a client and a viewer. The client does the number crunching, and the viewer shows off a graph of a protein (Figure 2). To fetch the software, pick up the deb or rpm file at the Folding@home website [3]. Install with

sudo dpkg -i fahclient_7.5.1_amd64.deb

or

fahviewer-7.5.1-1.x86_64.rpm

Figure 2: Folding@home viewer showing a rendering of a protein.

Space Race

Many research projects have not found a good way of classifying their data. The Galaxy Zoo project has invited volunteers to classify pictures of different galaxies. Programs for these classifications has proven notoriously difficult to write, so in this case, you will be donating your personal brain power.

NASA also has a software package for analyzing images from their missions; you can download the package from the US Geological Survey website [4].

The package is for Ubuntu 14.04, which means your libraries are probably too new, which could cause problems with a more current release.

Use a virtual machine to get started; osboxes.org has 14.04 appliances. They also support Fedora 25 and Debian 8; the source is available on GitHub [5].

The complete data set is 130GB; if you think that is a lot, you need to choose specific missions to work with. If you use the command line, the server expects the rsync command. If you want to follow an active mission, you can just rerun rsync on a regular basis, and the command will download only the difference.

The following command downloads the data from the Cassini mission.

rsync -azv --delete --partial isisdist.astrogeology.usgs.gov::isis3data/data/cassini data/

The Cassini data weighs in at around 14GB on disk. The data comes as pictures and other files that define where the spacecraft was at the time the picture was taken. To get accurate results for things like the height of a mountain, you need the position and angle of the camera.

With the data size in mind, consider using a SPICE server. Your software can use the SPICE protocol to pick up data on demand, instead of downloading it all in advance. Add -exclude='kernels' to your rsync command and run the following command:

spiceinit -GUI

You will now use spiceinit to create a specialized data file for your particular needs (Figure 3). If Cassini is not enough, you also have 9 missions of the European Space Agency.

Figure 3: Using spiceinit to create a data set.