Dave Cech of IDT asks whether it is time for silicon and software to drive new market opportunities

There is no doubt that silicon innovation has been critical to the advancement of a variety of technology-based applications over the past 30+ years, but it is not always clear whether the silicon is driving the advancement, or the application itself.  Silicon advancements come in a variety of different forms as well, as advances can be related to a new process technology, a new architecture, or integration of certain functions and features.

Silicon is sometimes purpose-built to fit a requirement, meet a standard, or advance a line of products to its next iteration.  Conversely, in other situations, designers take advantage of silicon innovations and advancements to enable a new product, application or feature.

The key to success is to have a silicon partner that understands your end market requirements and can identify performance bottlenecks before they occur, proposing new and innovative product architectures that deliver a solution to the problem in a different and/or more effective way. 

Silicon vendors also need to be in touch with the requirements of the end user.  Understanding what the end user wants and then empowering the OEM to deliver those solutions via new silicon architectures should always be the goal of a progressive silicon vendor!

One of the best examples of silicon’s ability to drive new markets today is the pending transition to IPv6 on a wide basis for IP based traffic.  IPv6 presents a great opportunity to expand the IP address space to enable a wide variety of next generation networking applications that require a rapidly increasing number of end points with unique IP addresses.  However, managing services in an IPv6 network is a greater challenge than with IPv4, as the increased IPv6 address size makes achieving acceptable network performance levels much more challenging. 

An IPv6 forwarding lookup, which defines where to send a packet, requires a 128-bit lookup based on its IPv6 destination address.  In contrast, the existing IPv4 standard requires only a 32-bit lookup.  IPv 6 policy lookups are even more of a challenge, as they are typically 296-bits long or more!  The increase in address size will place heavier demands on network equipments’ forwarding and policy evaluation performance, with functions such as quality of service (QoS), filtering rules for access control lists, and billing all presenting major challenges.  Advanced packet-processing silicon solutions, such as network search engines (NSEs) with wide search widths and high performance, will prove to be key components of the IPv6 solution.

As a silicon vendor, recognising that IPv6 lookup challenges exist is certainly a key to success, and in turn, being able to build a solution that provides an order of magnitude improvement in price and performance is another.  The final step is to partner with cutting-edge system vendors and ultimately craft an application that will fully exploit the benefits of IPv6 enabled by innovative silicon solutions.

Dave Cech is director of marketing for IDT’s IP co-processor product line

More information can be found here.