A programming environment for packet-processing systems

Vin, Harrick M.; Mudigonda, Jayaram; Jason, Jamie; Johnson, Erik J.; Kunze, Aaron; Lian, Ruiqi

doi:10.1016/b978-012088476-6/50009-5

Cited by 7 publications

(4 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NOVA has features for specifying header representation, including a layout construct for packed and unpacked forms and an overlay construct for alternative organizations within a layout. Network Protocol Description Language Vin et al proposed the Baker programming language [14], a C dialect augmented by data-flow concepts, which emphasized pipelining packet data from function to function.…”

Section: Related Workmentioning

confidence: 99%

Using Packet Processing Object Modules Interchangeably as Stand-Alone Programs or “Multi-app” Components

Duncan

Jungck

Ross

et al. 2014

Int J Parallel Prog

View full text Add to dashboard Cite

The basic model for processing packets and routing them from one hardware system to another can be extended to describe virtual packet transmission from one module to another within a single program image. Representing virtual and physical ports interchangeably and embodying their routing relations in runtime data structures makes it possible to produce object modules that can be deployed in a variety of roles without recompilation. We describe a fielded system that combines (a) application virtual ports, (b) a software tool for specifying packet routing to virtual and physical ports, (c) an advanced linker that encapsulates routing actions in runtime tables within a single multi-application image, and (d) a table-driven dispatcher that effects control flow and packet routing. We present results that show performance advantages and demonstrate the usability and security benefits of reconfiguring and redeploying programs in object module form.

show abstract

Section: Related Workmentioning

confidence: 99%

Using Packet Processing Object Modules Interchangeably as Stand-Alone Programs or “Multi-app” Components

Duncan

Jungck

Ross

et al. 2014

Int J Parallel Prog

View full text Add to dashboard Cite

show abstract

“…The table in Figure 3 shows the mapping and start, finish times of the tasks using the list scheduling algorithm [10], the optimal solution and the randomized rounding (RR) technique based solution. In case of the list scheduling algorithm, task b will be mapped to the CP (since, communication cost(4) + processing cost on CP(2)=6 < processing cost on GPP (12)). Thus the greedy list scheduling technique will be trapped in a local minimal point.…”

Section: Figure 3 Dag and Mapping Solutionsmentioning

confidence: 99%

“…The first definition can be used before the system is designed and the second definition can be used after the system is designed. A profile-guided automated mapping compiler was developed for runtime performance enhancement in [12]. In addition to [10], this may be viewed as an example of the second definition.…”

Section: Figure 3 Dag and Mapping Solutionsmentioning

confidence: 99%

Resource mapping and scheduling for heterogeneous network processor systems

Yang

Gohad

Ghosh

et al. 2005

Proceedings of the 2005 ACM Symposium on Architecture for Networking and Communications Systems

View full text Add to dashboard Cite

“…the IXP has ME next-neighbor registers or Scratch Memory rings for this purpose). The most promising uses of runtime reconfiguration being explored include dynamically turning ME cores off to reduce power consumption when network traffic is low and dynamic code reconfiguration to adapt to changes in processed packet characteristics [36]. This part of Shangri-La is under development and not the focus of this paper.…”

Section: Runtime Systemmentioning

confidence: 99%

Shangri-La

Chen

Li²,

Lian

et al. 2005

SIGPLAN Not.

Self Cite

View full text Add to dashboard Cite

Programming network processors is challenging. To sustain high line rates, network processors have extremely tight memory access and instruction budgets. Achieving desired performance has traditionally required hand-coded assembly. Researchers have recently proposed high-level programming languages for packet processing, but the challenges of compiling these languages into code that is competitive with hand-tuned assembly remain unanswered.This paper describes the Shangri-La compiler, which accepts a packet program written in a C-like high-level language and applies scalar and specialized optimizations to generate a highly optimized binary. Hot code paths identified by profiling are mapped across processing elements to maximize processor utilization. Since our compilation target has no hardware caches, software-controlled caches are generated for frequently accessed application data structures. Packet handling optimizations significantly reduce perpacket memory access and instruction counts. Finally, a custom stack model maps stack frames to the fastest levels of the target processor's heterogeneous memory hierarchy.Binaries generated by the compiler were evaluated on the Intel IXP2400 network processor with eight packet processing cores and eight threads per core. Our results show the importance of both traditional and specialized optimization techniques for achieving the maximum forwarding rates on three network applications, L3-Switch, MPLS and Firewall. , program partitioning, dataflow programming.hand-coded assembly. Networks have mostly relied on the widespread use of a few core packet programs. To achieve high line rates, though, a network program usually has very tight memory access and instruction count budgets. In the past, careful handoptimization of assembly code has been the most effective means of achieving the required performance given the small kernels and difficult resource constraints.As networks have evolved, so has the code running them, becoming larger, more diverse and complex. More and more network protocols are being developed for specialized applications (e.g. wireless, VoIP, proxies, network-attached storage). Enhancements to base protocols have been implemented to satisfy load balancing, security and reliability requirements. Shipped network hardware must operate correctly in an increasing number of different configurations.Hand-coded assembly has become a hindrance to developing new network applications and updating existing applications. Reusing common routines written in assembly in different contexts is difficult, debugging assembly code is extremely tedious and maintaining assembly code is a time-consuming effort. Even stateof-the-art tools that abstract some of the assembly programming details still expose to programmers the multi-threaded packet processing cores, heterogeneous memories and custom communication topologies found on network processors.Shangri-La, which consists of a programming language, a compiler and a runtime system, simplifies development and accelerates perfor...

show abstract

A programming environment for packet-processing systems

Cited by 7 publications

References 27 publications

Using Packet Processing Object Modules Interchangeably as Stand-Alone Programs or “Multi-app” Components

Using Packet Processing Object Modules Interchangeably as Stand-Alone Programs or “Multi-app” Components

Resource mapping and scheduling for heterogeneous network processor systems

Shangri-La

Contact Info

Product

Resources

About