Home Documentation Forum Tracker Download

Welcome to Cumulus4j!

Data privacy protection: the grand challenge for cloud software

In this day and age, theft of highly classified data is a real threat. For example, there have been several reported cases where sensitive banking information were stolen and subsequently sold to tax authorities. Or, evidences in the media pointing to incidents where malicious hackers were able to get away with pilfered copies of huge databases from multinational companies. It is understandable that customers are worried about the security of their data and, as a result, most companies refrain from transmitting business-relevant data over the Internet to a service provider for outsourcing operational services (like invoicing or accounting).

One of the most important barriers for the success of business-operational services via the Internet - software as a service - is thus the lack of data security which would guarantee that not even a system administrator can dump and sell data.

Encryption for maximized data privacy protection

To overcome this barrier, Cumulus4j provides a secure architecture for trustworthy accounting software in the Internet. The demands on the trustworthy database made by an accounting software are naturally very similar to those of other enterprise applications. Therefore, Cumulus4j is an interesting solution to a multitude of applications.

In order to provide a trustworthy data management, all data must be stored in an encrypted form in the cloud. Though this can already be achieved by a disk encryption software (e.g. LUKS or TrueCrypt), this solution requires the hard disk to be mounted onto the operating system before any application can operate on the data. And as soon as the disk is mounted, an internal adversary (e.g. a system administrator) can easily grab all data via the file system.

Overview

Cumulus4j solves this problem and fulfills the following requirements:

  • Security: State-of-the-art encryption algorithms protect your data. All data are decrypted in the volatile memory only. There is - unlike when using disk encryption - not only one single key for all data, but multiple ten-thousands of keys. Additionally, the en-/decryption is done inside the application and not as a drive on the operating-system-level. This makes it nearly impossible for an internal adversary to steal more than a small number of data records (by RAM-dumping).
  • Easy integration: In order to ease the integration of Cumulus4j into an application, it has been designed as a plug-in to DataNucleus. This way, modern persistence APIs like Java Data Objects (JDO) or Java Persistence API (JPA) are made available, which facilitates the almost transparent integration of Cumulus4j into a cloud application (only very few lines of Cumulus4j-specific code are necessary for the key management).
  • Query-ability: Though all records are stored in encrypted forms, they can be queried efficiently via the query languages JDOQL and JPQL.
  • Portablility: Cumulus4j can be used with many common databases as backend (e.g. MySQL, Google BigTable, Hadoop HBase, MongoDB, etc.). It has been written in Java and runs both in OSGi and in classic environments (JavaSE, JavaEE).

The Cumulus4j project is sponsored by the German Federal Ministry of Education and Research (BMBF) within the scope of the initiative KMU-innovativ.

Cumulus4j is free software under the conditions of the AGPL.

Documentation
About
Project Documentation
Babel
Releases