Bridging the Software/Hardware Forwarding Divide
نویسندگان
چکیده
As network speeds increase, and network requirements diversify, networking hardware vendors are being pressed on two distinct fronts. To remain competitive, they must continually improve their cost/performance ratio at a pace faster than Moore’s Law. Over the past fifteen years, enterprise network speeds have increased by three orders of magnitude, from 10Mbps to 10Gbps, requiring that the cost of a unit of bandwidth (a gigabit port, or Gport) drop precipitously. At the same time, networks are expected to address myriad application, management and security requirements that were unheard of fifteen years ago, creating pressure to develop network forwarding devices that can flexibly accommodate these new demands as they arise. This trend is reflected in the many “open” platform initiatives launched by system vendors [10, 16, 22]. Despite significant research and engineering efforts, the industry has failed to simultaneously achieve low cost/performance and high flexibility.1 There are three basic approaches to building network forwarding devices currently pursued in the field and, as we relate below, they offer quite different tradeoffs between these two goals. In the commercial world, where cost concerns dominate, the most prevalent approach to building network forwarding devices (this includes both switches and routers, but hereafter we will use the term “switch” as shorthand) involves embedding the basic forwarding logic in ASICs. There are commodity forwarding chips (from, for example, Broadcom, Marvell, and Fulcrum) that cost as little as $20/10Gport, a pricepoint that would have been unthinkable only a few years ago. Multi-layer switches using these commodity chips support 240Gbps (24x10Gbps) of capacity for under $100/10Gport, and 640Gbps (64x10Gbps) chipsets will be generally available this year which should lower the cost even further. However, one cannot modify the hardware forwarding algorithm without respinning the chip. This forces packet forwarding to evolve on hardware
منابع مشابه
The DECNIS 500/600 Multiprotocol Bridge/Router and Gateway
1 Abstract The DECNIS 500/600 high-performance multiprotocol bridge/router and gateway are described. The issues affecting the design of routers with this class of performance are outlined, along with a description of the architecture and implementation. The system described uses a distributed forwarding algorithm and a distributed buffer management algorithm executed on plug-in linecards to ac...
متن کاملFlow Labelled IP: Connectionless ATM Under IP
IP traffic on the Internet and private enterprise networks has been growing exponentially for some time. This growth is beginning to stress the traditional, processor based design of current day routers. Switching technology offers much higher aggregate bandwidth but presently only offers a bridging solution. Various proposals are under way to support IP routing over ATM switching technology; h...
متن کاملService Characterization for Virtual Routers
In the past decade the virtualization paradigm has been remarkably successful in the server domain, and significant benefits are expected from the virtualization of network resources. However, to date the virtualization capabilities of proprietary systems remain limited while the impact of executing open software routers within system virtualization platforms has not been fully investigated. In...
متن کاملPossible, Probable and Preferable Futures of the Digital Divide
The digital divide is widely recognized as a contemporary problem between society and techno logy. Strategies for bridging the digital divide are often informed and guided by quantitative assessments of the deployment of information communication technologies. There are few rigorous qualitative attempts to assess the digital divide from either an ethnographic or a futures-oriented perspective. ...
متن کاملDesign and Evaluation of Cloud Network Optimization Algorithms. (Optimisation des réseaux virtuels : conception et évaluation)
The rapid development of software virtualization solutions have led to huge innovation in Data Center Networks (DCN), with additional capabilities to perform advanced network functions via software elements. Novel protocols designed in the recent years allow for a large path diversity at the edge server level by means of multipath forwarding protocols and virtual bridging functions. Moreover, i...
متن کامل