Middleware, development tools, realtime operating system
software and services for superior embedded design


Home
Download Center

Lowering the Development Costs of Network Elements through Software Partitioning

QNX Software Systems
Product downloads
QNX SDP 6.5.x
Products + Updates
BSPs

General
Benchmarks
Online tutorials
Pre-release software
Porting libraries
Third-party software

Collateral
Whitepapers
Brochures
Market briefs
Partner briefs
Product briefs
Technical articles
Product docs
Japanese docs
File Information Parent Folder
Download Now
Lowering the Development Costs of Network Elements through Software Partitioning
February 22, 2007
By providing CPU guarantees for each software subsystem, partitioning can dramatically simplify software integration efforts.
Register

  File Information  
Filename:qnx_partitioning_for_networking_whitepaper_RIM_MC411.52.pdf
Size:167.12 KB
Classification:
Check Sum: 1341229769 228568
MD5 Sum:0d06085516fd2da55a594f54d8247f00
  Additional Notes  
Networking and telecommunication OEMs are at the forefront of design complexity. A network element can contain hundreds of thousands or even millions of source lines and employ hundreds of software tasks, all of which contend for a finite amount of memory and CPU time.

To build these massive software systems in the least possible time, OEMs typically opt for parallel development, dividing the work among multiple project teams. Unfortunately, the parallel development paths invariably lead to performance issues at the system integration phase, when, for the first time, subsystems built by the various teams begin competing for system resources. Subsystems that behaved correctly in isolation now respond slowly, if at all. In many cases, these resource-contention problems emerge only during integration and verification testing, when the cost and schedule impact of software recoding is at its highest.

Download this whitepaper and learn more about how to avoid weeks or months of corrective actions resulting in increased costs and negating the time-to-market advantages of parallel development.