Slashdot Mirror

I'm not talking about T1 repeaters; I'm talking about DSLAMs. And DSLAMs are not tiny things -- not the ones with enough ports make them worth the expense. I have a spreadsheet around here that lists (listed... it's dated 2004) every installed DSLAM in bellsouth's network. (hey, they put it on the web...) I don't recall seeing any "tiny" line powerable hardware (~30W) in that list. Basically, because Bellsouth isn't going to waste their time installing a DSLAM for 24 lines. (Zhone makes one that's 48 ports, but that's still too small.) 26AWG can safely provide max. 15W to ~12k ft. (depending on how much voltage the local laws permit.) [Pedestal Networks did the math for me.] Almost all of BS's wiring is 26AWG.

I'll be sure to look at what's on the corner next time I head to the grocery store. In theory, there's one right there, but I think Google is about 100ft too far to the right. However, at 1am I don't think I'll be stopping to take any pictures :-)

Wow,

Finally, A linux subject that the slashdot crowd is silent about. Since it references debian the ubuntu comments were inevitable.

In reality this goes far beyond a debian discussion and is actually great news for the business of linux. In general Carrier Grade OS's is a way of saying that the OS used for certain carrier/telecom applications must follow a certain set of rules and standards.

This is important because carrier grade linux has to support a ridiculous feature set in order to achieve guaranteed uptime status. The more distributions there are with these features the more they become into mainstream kernels.

Another key advantage is that most carrier grade OS's charge you an arm and a leg such as montavista, sun and others. This prevents the smaller players from really being able to get into the carrier business and offer services because of the exorbitant costs of such applications. However this doesn't really figure into most of the current American telecom companies but where it really does make an impact is in third world/developing countries. The telecom's there are either government run or private institutions that can only offer their services to the wealthy.

Combine a platform like this with an open source SIP server like http://www.sipfoundry.org/ and you have the makings of some lower priced and more ubiquitous VOIP solutions for developing countries.

I think that steps such as these are what will make linux such a key component in other countries where the infrastructure is not yet built and can be based upon the matureness of open source technologies.

Here is my synopsis of what CGL entails. The full article can be found here http://www.commsdesign.com/design_corner/OEG200208 27S0008

Overview
--------
The OSDL CGLWG (Carrier Grade Linux Working Group - whoo thats a mouthful) defines three main types of applications that carrier-grade Linux will support -- gateways, signaling servers, and management.

Categories
----------
To build an effective specification for the applications above, the CGLWG is working on seven categories: standards, platform, availability, serviceability, tools, performance, and security. Work within each of these categories is broken down into three priority levels: level 1 defines first release requirements, level 2 defines second release requirements, and level 3 defines future release requirements.

Category 1: Standards - IPV6, Posix etc etc

Category 2: Platform Requirements - Hot swappable devices, uptime, kernel hardware error correction blah blah

Category 3: High Availability - Hmm pretty self explanatory

Category 4: Serviceability
On the serviceability front, the CGLWG is currently trying to define requirements for resource monitoring, kernel crash dump and analysis features, structured kernel messages, dynamic kernel probing, hardware error logging, and remote access to the event log.

The best feature of CGL in my opinion

Category 5: Tools
Priority 1 in the tools category will focus on the development of debugger support for threaded programs, kernel debugger support, and a kernel crash dump analysis tool.

Second best feature

Category 6: Performance
In the performance category, priority 1 performance features required include millisecond real-time (less than 10-ms worst case latencies), pre-emptible kernel, RAID 0 (striping), application pre-loading, and a scaling analysis and report to identify scaling bottlenecks.

Category 7: Security
The current version (V1.0) of the carrier-grade Linux specification does not have any requirements for security, but this will be addressed in a future version of the document. Security is recognized by the CGLWG as a key component of high availability.

YIKES!!

So there it is CGL 101. It will enable small development teams to offer real solutions in the behemoth telecom field.

-------------

For info go see: http://www.commsdesign.com/showArticle.jhtml?artic leID=177105035

That's pretty hard to do since GC isn't deterministic--standard implementations of new/delete aren't deterministic either but C++ allows you to override them and there are deterministic algorithms available.

It it a lot less hard than it used to be, because a lot of work has been put into Java garbage collection to prevent it having an impact on performance. The possibility of using Java for real-time work was realised some time ago. Here is an article from 2001:

http://www.commsdesign.com/design_corner/showArtic le.jhtml?articleID=16503463

I am not an expert in this area, but I have seen demonstrations of how effective this can be. In 2004 Sun even demonstrated real-time device control and non-real time general Java apps running on the same VM.

My personal view is that garbage collection is now finally mainstream, and manual management of memory will be eventually relegated to niche uses.

It's not disinformation. It's just out of date. Originally, the MPEG-4 licensing was to be based on the duration of the video.

Group revises MPEG-4 licensing terms

"Samsung is currently shipping 4 million 3-Mpixel camera modules a month ... The company's new camera module, which groups the 3-Mpixel sensor with LEDs and optical devices, takes advantage of Samsung's packaging expertise, Moon said. The module uses a "liquid-lens" mechanism to obtain a 2x to 3x optical zoom. The liquid lens is an electrolytic oil-and-water combination, a convex bubble, which changes its focal length when given a variable voltage. The response time is about 75 milliseconds."
http://www.commsdesign.com/news/tech_beat/showArti cle.jhtml?articleID=52600845

I doubt they'd do this without investigating patents.

No, as I mentioned in my original post, phased arrays can add a spatial dimension (or two, or three) to the detection of transmitted signals, and therefore to the detection of transmitters. Another example is stereoscopic vision - using parallax among multiple detectors allows resolving distinct simultaneous transmitters on the same frequency from each other, by their spatial separation. So you can distinguish between the red stoplight up ahead, and the red brake lights dead ahead, while you're driving. Or two different WiFi stations transmitting from opposite corners of a street, even on the same frequency.

As for practicality and elegance, Vivato is one vendor that has been producing business advantages with their WiFi phased arrays.

Well maybe not in the US as a whole, but it is occuring in certain municipalities like Provo and Spokane.

In other words, assuming *zero* processing overhead, we're 25 MIPS short for wire-speed encryption.

These are very rough numbers, but think of it this way: do you think Cisco (or whoever) spec'ed a processor substantially faster than what they needed? From my peronal experience, embedded processors do not usually have more than a few percent more performance than they need: rarely do they have even 30% more performance than they need. Even if they design a system with a way-fast processor, one of two things happen: their code bloats to use that speed (or they quit optimizing because they don't need to), or they end up buying a lower-cost, slower processor for production!

In short, it's highly unlikely that the Wrt54g will have anywhere near the CPU power to do wire(less)-speed AES at 54Mbit. Half that? Maybe, but not all of it.

Your reference for this? The first few references I found on Google all disagree with you:

• the human eye is more sensitive to luminance than chrominance
• We know that the human eye cannot perceive differences in color as well as it can differences in intensity.
• The human eye is more sensitive to brightness (gray scale data) variations than to hue variations
• the human eye is much more sensitive to luminance than chrominance
• the human eye is much more sensitive to brightness variations in gray-scale than to color variations

However, cell phones are an incredibly restrictive environment

Perhaps you can think of ways that cell phones are restrictive, but here's a headline from the front page of Today's EETimes.

GSMA chiefs stress open standards as phone users top 1 billion

We used the AT&T DSP32, a 12.5MFLOPS DSP, 15 years ago at Array Technologies. Programmable in a native C source code, with multiply-accumulate (MAC) instructions optimized in microcode, the DSP32 was lightning fast at y = mx + b equations in its arithmatic logic unit (ALU), and its control logic unit (CLU) was also very fast at branching, including no-overhead looping. Linux runs on one of its many fascinating descendants, the Xilinx Virtex-2 Pro.

A good introduction to the technology and how it differs from RF is here

Building multiprocessor chips, or chips from arrays of processors has become a fairly hot design approach. There are a number of companies using it. It seems to be especially popular in the reconfigurable computing area. There is an interesting paper here. These processors go well beyond the current crop of dual CPU core chips like the P4, Power 5, and Ultrasparc IV.

Clearspeed's chip is a static 64 processor array chip aimed at FPU intensive applications, but there are many more things that you can do with array designs.

Mathstar is building a reconfigurable chip with hundreds of elements availble in various mixes of processors, memory blocks and other components. They are trying to replace ASICs and FPGAs as a platform for some part needs. There was a story on their architecture in EE Times a couple of months ago.

Intel is wokring on an array based processor aimed at the radio / communications market. I will be interested to see if their work with these chips ends up being used in other Intel chips. That could be deadly. So, the Pentium-X sucks at that task today? [Morph] Not now!

Phillips has what they call Silicon Hive technology which is another reconfigurable processor of functional blocks.

There have been plenty of companies using arrays and reconfigurable techniques too, like Altera and Chameleon.

Sun bought up a start up and is developing massivly multithreaded processors based on the start-up's technology. They call it Throughput Computing. They claim that in about two years they will have a chip 30x faster than todays designs. I'll be very interested to see if they can do that.

The next couple of years should be very interesting on the processor front.

The MIPI spec that is developed by this consortium is also mentioned here. They mention ST as a fourth player as well.

I didn't think PCI-E and hypertransport were in competition, eg: How HyperTransport and PCI Express complement each other. Could you explain why intel chose one over the other?

http://www.commsdesign.com/story/OEG20030423S0042

Unfortunately this is a rehash of old news. The fact that Mobiles, PDAs and Laptops *can* cause interference has been widely known for a long time. Anyone that has flown in the last 5 years will be familiar with the warning to turn off these devices on take off and landing. The possible Ban on laptops etc relate to the introduction of 'ultrawideband' capabilities for these devices which 'could affect a plane's electronics, including its instrument landing system and its collision avoidance systems'. Ultrawideband devices are expected to hit the stores this year, and will range from laptops to PDAs to the following military applications: Since Aircrew will not be able to tell the difference between UWB devices and regular laptops, it seems that a blanket ban may be applied. A good overview of Ultrawideband and its political consiquenses can be found here An article on Ultrawideband and its effects on aeroplanes can be found here

For those whom are not aware of this technology, I gladly provide the following linkage:

-- Article @ Network World Fusion

-- Article @ Comms Design

-- Published Standards & Drafts

Enjoy! :-)

is available on commsdesign.com

What Intel is saying:
IEEE 802.16 spec could disrupt wireless landscape

No, here it is: http://www.commsdesign.com/story/OEG20030130S0055 Maybe I'll click HTML instead of Plain old text next time.

one, two (looks interesting), three (looks interesting and authoritative).

ooops for got to paste the link! 802.11 / telecoms IP story.

> Both CDMA and GSM/GPRS cost about the same to roll out.

Yes, right. But, it will cost *existing* GSM operators a lot to roll out GPRS on their existing networks. And when they go to 3G, it'll cost a lot again. CDMA is a bit more reasonable: you can roll out higher-speed services incrementally.

Ask yourself this: if GPRS is so easy to roll out, why has not one GSM provider in India provided GPRS so far? They've been talking about it since August 2001, it is in danger of turning into vaporware!

So the only way to build a high-speed GSM network is to build a 3G network from scratch. See my other reply in this thread -- people have tried this (notably DoCoMo in Japan) and have had egg on their faces.

Check out this new antenna design that supposedly gives 802.11 a three-mile range.

Sun acquired Pirus Networks to help them on a chassis with FibreChannel, iSCSI, and perhaps InfiniBand.
http://www.byteandswitch.com/document.asp?doc_id=2 1423

Before that they picked up Dolphin Interconnect to help them make a 4x (30 Gigabit/sec) InfiniBand Host Channel Adapter.

Here is an article from an EETimes Network site, CommsDesign with some details.

http://www.commsdesign.com/news/tech_beat/OEG20020 919S0076

It is definitly interesting stuff. Everyone is trying to do Shared I/O and I/O Virtualization; maybe Sun can get it right.

EETimes has a sorter article, and Comms design has a more in-depth article covering some of the problems TI may face.

Most people use a 4 chip solution - with each chip's process suited for its use:

- power management (high current)
- baseband/applications processing (good routing)
- memory (high density)
- RF/IF plus power amp (high speed, high voltage)

How expensive/feasable is it going to be to put a high-density ferroelectric EPROM memory along with SDRAM and a 6-volt RF power amp?

A slightly more well known source reports on the TI GSM chip.

TBQH, I think the article is correct in saying that commercial WiFi hotspots are unlikely to stick around for the long haul. A few specialist "captive audience" spots might work out, and in the short term the novelty factor might support coffee-shop spots for a while; but long term they're screwed when 3G rolls out. Right now though, WiFi is still at the top of my shopping list...

WiFi itself will carry on. The fast roll-out for corporate networking, the convenience for home networking, and the opportunities for community networks all count for a lot with me. Much more so than (yuk) Bluetooth. I think 802.11a/b formats have a rich future ahead of them.

Me, I want to play with 802.15.4 aka ZigBee. Sure, it doesn't have the ethernet-sized bandwidth but it's really cheap, low-power, and thinking about the possibilities of swarms of tiny ZigBee-enabled devices makes by brain hum. In a good way. :}