|
Summary
Use/Applications
Technical
Pocket Soft's Bona Fides
More Information
|
Technology
- New generation byte-level differencing designed for use by F1000 and Government.
- Windows, Linux, Unix
- Native 32-bit, Native 64-bit
- Product Includes: differencing engine, client side apply software with source code and SDK
- Lossless compression, 100% identical restoration
- Compression ratios depend on the number and pattern of changes. Typical values are 90-95%. Values of 99% are not unusual.
- dfc-gorilla may be used in multiple simultaneous instances without adversely affecting the running operations of the system. Previous byte-level difference algorithms were designed to run as a single instance on a dedicated machine.
- Linear build time: The time to identify and encode the byte-level differences between files is linear with respect to input file sizes. Previous byte-level difference engines required quadratic runtimes.
- Sequential file access:
dfc-gorilla uses a single read pass during data comparison, with no backward seeks. Ideal for situations that involve costly seeks (tape media), or in cases where back-seeking is not possible (streaming), this file access approach is unlike any previous byte-level difference technology.
- Memory and other system resources are limited and definable.
All of the resource requirements of dfc-gorilla are independent of the sizes or numbers of files upon which it operates. In addition, the actual amount of various resources may be defined. As little as 32-64MB of RAM and no temporary disk space, is required.
- Memory is limited and chosen by the developer: Typically 32-64MB. Can be as little as 16MB. Previous byte-level difference engines required memory equal to 1.5x the size of the old and new files combined. When the requirement exceeded available memory, a swap to disk or unrecoverable error would occur.
- Other resources: Previous byte-level difference engines required temporary disk space and swap file space. dfc-gorilla requires none.
|