Here is a white paper sponsored by SonicWALL talking about the advantages of Multicore Unified Threat Management (UTM).
Abstract:
“Network communications no longer just rely on store and forward applications like e-mail. It has now grown to include real-time collaboration tools, Web 2.0 applications, instant messenger (IM) and peer-to-peer applications, Voice over IP (VoIP), streaming media [...]

[Read more →]

In the first part of this “Is the Free Lunch Really Over?” series, Michael Wrinn wrote about view of Amdahl’s law and scaled speedup (of John Gustafson) of many-core processors for Intel Software Network. In the Part 2 of this series, he discusses the efficient use of locks, and in particular at two guiding principles: [...]

[Read more →]

Here is the paper about Larrabee processor that provides more details into its hardware architecture and scalability studies.
Larrabee: A Many-Core x86 Architecture for Visual Computing [Weblink from Intel]
Larry Seiler, Doug Carmean, Eric Sprangle, Tom Forsyth, et al.
SIGGRAPH 2008

[Read more →]

Multicore mania has definitely hit the embedded networking market, but as the dust begins to settle it has become clear that many important architectural details need to be examined closely before decisions are made about how to partition applications across multiple cores. Stephan Cole et al. from Freescale write an article on how to partition. [...]

[Read more →]

The traditional approach to application performance was to simply wait for the next generation of processor. Most software developers did not need to invest in performance tuning, and enjoyed a “free lunch” from hardware improvements. With the industry shift to multiple core systems, the situation has changed: performance has to be realized through concurrency, with [...]

[Read more →]