Sharing and repeating scientific applications is crucial for verifying
claims, reproducing
experimental results (e.g., to repeat a computational experiment described
in a publication), and promoting reuse of complex applications. The
predominant methods of sharing and making applications repeatable are
building a companion web site and/or provisioning a virtual machine image
(VMI). Recently, application virtualization (AV), has emerged as a
light-weight alter- native for sharing and efficient repeatability. AV
approaches such as Linux Containers create a chroot-like environment, while
approaches such as CDE trace system calls during application execution to
copy all binaries, data, and software dependencies into a self-contained
package.
In principle, application virtualization techniques can also be applied to
DB applications,
i.e., applications that interact with a relational database. However, these
techniques treat a database system as a black-box application process and
are thus oblivious to the query statements or database model supported by
the database system. To overcome this shortcoming, and leverage database
semantics, we have introduced light-weight database virtualization (LDV)1
is light-weight as it encapsulates only the application and its necessary
and relevant dependencies (input files, binaries, and libraries) as well as
only the necessary and relevant data from the database with which the
application interacted with. LDV relies on data provenance to determine
which database tuples and input files are relevant. While monitoring an
application to create a package we incrementally construct an execution
trace (provenance graph), that records dependencies across OS and DB
boundaries. In addition to providing a detailed record of how files and
tuples have been produced by the application, we use it to determine what
should be included in the package.
The primary objective of this demonstration is to show the benefits of
using LDV for repeating and understanding DB applications. For this, we
consider real-world data sharing scenarios that involve a database and
highlight the sharing and reproducibility challenges associated with them.
We give an overview of our LDV approach to show how it can be used to build
a light-weight package of a DB application that can be easily shared and
reproduced. During the demonstration, the audience will experience three
key features of LDV: (i) its ability to create self-contained pack- ages of
a DB application that can be shared and run on different machine
configurations without the need to install a database system and setup a
database, (ii) how LDV extracts a slice of the database accessed by an
application, (iii) how LDV’s execution traces can be used to understand how
the files, processes, SQL operations, and database content of an
application are related to each other.
1 https://github.com/legendOfZelda/LDV.git, a tool for creating packages of
DB applications. An LDV package
