Intel's Communications Effect A great article in Fortune on Intel's investments into greater manufacturing capacity and Digital Signal Processors (DSPs).
"...the telecom industry has never really benefited from Moore's Law because most of those proprietary network processors weren't made in large enough volumes to warrant leading-edge manufacturing technology. A standard part made with state-of-the-art manufacturing technology, coupled with a set of standard programming tools that are derived from PC software tools, could finally bring Moore's Law to telecom," said Sean Maloney, executive vice president for the Intel Communications Group.
Actually, Telecom has achieved Moore's Law in optical transport; price/performance gains below Layer 3 that accellerated the supposed bandwidth glut. This helped foster a new traffic engineering paradigm, achiving service quality through over-provisioning capacity (vs. centralized intelligent management of channelized capacity).
If you assume Intel brings Moore's Law to Layer 3 and up, it would change the way traffic is engineered. It might bring down the hop-tax for traffic. It might allow scalable intelligence to be applied at edge devices without the cost of complexity. It might allow quality to be guaranteed through techniques that have yet to be widely deployed like MPLS. All while decreasing the cost of boxes that are the vast majority of carrier CapEx.
The problem is telecom CapEx is that the markets are not going to allow capacity in excess. Smart and faster builds according to pre-sales will be required. And over-provisioning capacity to assure quality will be called into question. If Moore's Law brings down the cost of boxes and systems houses like Cisco enable more granular upgrade paths, a more variable cost structure will emerge that can be accepted by the markets.
But the problem with the service provider business model isn't just CapEx, its OpEx, namely people and billing. SG&A costs that average at 25% of revenues, the highest of any industry. And billing costs at 20%. If telecom is to become sustainable, software solutions to this problem are the key.
Perhaps looking to the IT industry's experience with Moore's Law provides insight into the changes to come. When Moore's Law drives the physical commoditization of Datacommodities (storage and processing in IT), first as hardware solutions (PCs) and then as services (utility computing), players change roles. Intel increasingly takes on the functions of system houses (IBM, Sun, HP). Systems houses move into software (data center virtualization, utility computing, autonomic computing). Software providers move into services (IBM's acquisition of PWC consulting, HP/Compaq merger driven by services).
If Moore's Law drives the physical commoditization of the other Datacommodity, bandwidth, it will accelerate a similar shift. Intel will increasing fabricate the engineering functions previously provided by system houses like Cisco, Lucent and Nortel. As network elements are built upon a defacto standard set of components, they too will trend towards commoditization. Increasing intelligence in network elements will allow greater virtualization for management. As margins errode in their core business, systems houses will further standardize element interface and open their operating systems to eliminate their barrier to entry OSS (proprietary element management) [admittedly this is the weakest link in the scenario]. Systems houses will be well positioned to offer new network management platforms and move up the value chain. Virtualization and autonomic techniques will drive down cost of people and billing. Software providers will increasingly move into services (similar to what CDN & VPN providers have done).
Regardless of the value chain shift, the move towards increasingly intelligent edge devices at lower cost would enable new revenue models. MPLS engineered traffic that is service and application specific with QoS guarantees between service providers (autonomically negotiated and settled) could offer a yield management driven revenue model. Making these network services available at the edge, for access and web services, enable growth at the edge.
Other blogs are highlighting how this effects wireless broadband: Steve Stroh, John Robb, Kevin Werbach. And Brad Delong says "In the future it sees, it destroys several industries--and creates entire new ones"
10:25:33 AM 
|
