We plan to have 3 major releases every year, with intermediate minor releases adding functionality to the aforementioned major releases. We typically number our releases based on the following rule: x.y.z; here, x.y indicates a major release with a significant advancement in the software’s capabilities (y increments from 0 through 9, before there is a rollover into x); z indicates improvements to the x.y release which is typically the result of bug fixes.
The Sustain software stack contains the following sub-projects. Our code base has been developed in several languages including Java, Scala, Rust, C++, Python, and Javascript.
| Aperture | Visualization of spatiotemporal data based on sketches, shape files, observations, and images |
| Stash | Distributed caching for spatiotemporal visualization |
| Sustain DHT | Distributed hashtable based storage of sketches and strands. Leverages proxy arrays and LSM trees for scaling. |
| Atlas | Encompasses the HDFS-compliant distributed file system, and support for constructs such as RDDs, Datasets, and DataFrames to facilitate interoperation with analytical engines. |
| Synopsis | Sketching of spatiotemporal data. This also includes support for imputation of missing data at diverse spatiotemporal scopes. |
| Polyglot | Inter-operation with diverse encoding formats and storage frameworks. |
The software is open-source, and repositories are available at GitHub.com/Project-Sustain.