My first impression is that you're on your way to designing a machine-turned, anodized, high performance buggy whip. Sorry, but I don't believe wired is the future - it's ultimately hard to scale. I would think advances in wireless and wire guides will turn the whole question of switching points into a virtual one.

But if you insist on using wires or light pipes, go for the solutions that have the best packing density. And for solutions on the scale you're looking at (25 million ports - are we talking part of a large municipal broadband rollout here?) -- I'd suggest you try to contact your counterpart at Krone. Because it ain't all Cat-6 any more.

posted 8 months ago

Well, when it comes to wired vs wireless, both will have their place for the forseeable future. I can remember pundits saying that hard drives were dead many times, but anything larger than a low spec Netbooks still comes with a harddrive.

Certainly on the IP front we have large service providers coming to us for multiple X 10G ports for Internet connectivity. We don't however supply the residential market so anything about 10G to the home is beyond my ken.

Capacity to the corporate office is of course still growing, but the trend to host applications and content in the network/cloud is slowing that same trend. The issue is not getting 10GE into the office, or even distributing it, it is the cost of ensuring that the infrastructure (power, cooling, security) is available 24 X 7.025...

Wireless helps, but I can't use my bluetooth stereo headsets when walking in downtown London because of WiFi hotspot interference. If every office has 10Ge it had better be at a dedicated frequency, and wireless networks will probably require planning permission for a new building profiles along high capacity 10GE wireless routes.

The cost of capacity will decline, but the per mile/kilometre cost of installing fibre won't.

posted 8 months ago

Hi,

The answers depend on the architecture. I pretty much expect hardware prices to be price elastic. Hardware prices are not the issue, it is concomitant build volumes.

If you plan to configure 'en-mass' you need to take control over the management of the configuring software and it should be 100% automated.

I would not build my own switches - you need to partner: unless you are particularly clever!

posted 8 months ago

1. The cost will be around 1000$/Month since in 2005 one hundred(100Mbps) cost in japan 70-100$ so 1Gb will cost almost 100$ and 1G 1000$ after 36months

2. since the max uplink is actually 40G you can have 4 port of 10GE but I think you can have 2 redundant uplink for 8 GEs is still possible

3. The good supplier is always from the private sector (as much as I know) however you can find good one in publiic


4. Built your own switch is a must. Because the price will be cut at least 40 %. You will spread the technology and solution, and will increase your revenue. The most important is : you will be technology driver for manufactorers not follower (you will manufacture products based on requirements. Before operators await vendors to propose new solution based on their own roadmap and market/business control...)

posted 8 months ago

Hi Neal,

10GB is a very good start. Just look at the laptop / PC Market where Volumes of storage is supported. As the applications are getting complicated and the size is increasing exponentially there is an ongoing need for much needed bandwidth and QoS. Focus should be on "beyond 10G". With Web2.0, Video Sharing, etc.. the bandwidth will be fully consumed.

VBR/ Wallis Dudhnath

posted 8 months ago