17/06/2005

Home at OpenSolaris.org

Home at OpenSolaris.org

Sun have now made OpenSolaris available. There are still encumbered bits of Solaris in there, but they've made the tools available to build Solaris on Sparc and on x86/x64.

Anyone can get an account (you need to register). Solaris is a good operating system, whether this will make a dent in the Linux market is debatable, but since they're giving away Forte development tools (which produce highly optimised code for at least the Sparc architechture) it should attract quite a few developers.

| ARCchart | Blueprint : WiMAX hype becomes hysteria

| ARCchart | Blueprint : WiMAX hype becomes hysteria

Arcchart's view on why WiMAX hype has reached hysteria. Intel claiming broadband wasn't fast enough and WiMAX offering 100Mb/s was the answer.

Unfortunately what Intel don't say is that to achieve 100Mb/s you need about 30MHz of spectrum, which is difficult to get hold of. Most operators (say operating in 3.4 or 3.5GHz or in the UK even 5.8GHz) have maybe 10MHz - which supports about 30Mb/s payload.

30Mb/s isn't bad at all, but that's the bandwidth available from the basestation and it's shared between all the end users. Someone could offer it all to one customer, but then it's unlikely the economics wouldn't stack up. Offering 1Mb/s at 20:1 contention (equiv to BT's business ADSL service) that would support 600 users, at 5Mb/s it goes down to 120 and at 10Mb/s only 60 customers. Basestation equipment runs at £20K+ so the more customers you can support, the better.

Consider the wholesale price of a LAN Extension Service (LES) in Central London, about £6K install, and £1.2K pa. OK that's a point to point service with no Internet bandwidth, but it shows how wired connectivity pricing is dropping (and it can be upgraded to 100Mb/s or even Gb/s as it's fibre).

But that's not the end of the story, even taking DSL technology the ITU have just ratified VDSL2 which now supports 100Mb/s, it doesn't support long distances but there's already chipsets available. ADSL2+ is storming ahead and that offers 20Mb/s (short distances) but does better over long distances than VDSL(2), so most providers are opting for it.

Yes there may be a market for WiMAX to reach the areas DSL doesn't, but it's diminishing and the economics are not proven.

16/06/2005

metranet

metranet

Interesting service, offering a WiMAX network in the Brighton and Hove area. Not much technical flesh on the site, but seems they're offering 2Mb/s, 5M/bs and 10Mb/s symmetrical broadband access over wireless (contended at 10:1).

They don't say what equipment they're using nor what spectrum. Do Ofcom know about them?

Also if they were founded by the council, they may fall foul of state aid regulation?

Does look like an interesting system though.

14/06/2005

BT finally ready to launch Bluephone

BT finally ready to launch Bluephone

BT have at last got their act together and will be lauching their Bluephone technology. This is where a dual-mode handset use GSM most of the time, and roam on to a Bluetooth basestation in the home (or office). BT have actually launched this in partnership with Vodafone.

The Bluetooth basestation is connected to the customers' broadband connection and the voice call will be backhauled via IP.

Unfortunately calls to the handset will always go via GSM at normal Vodafone rates, but BT will offer landline rates for the outgoing calls via the Bluetooth basestation.

The technology is actually part of the UMA (Unlicensed Mobile Access) and anyone can download the specifications from the UMA Technology site. Maybe someone can provide an alternative basestation that is compatible with Bluephone.

Nokia: Skype must change or die - silicon.com

Nokia: Skype must change or die - silicon.com

Well there's a suprise, Skype has to find a business model or it wont survive? Isn't thast true for any business? Skype have a huge user base and their model could be to just sell out to another company?

Nokia are pushing standards based VoIP (SIP) over mobile data. That's all very well, but at the moment (in the UK at least) data charges are rediculously high on most networks, 3 don't even offer a data service to the Internet (walled garden).

Even if data charges drop dramatically, they'll have to drop a long way to get close to bundled telephony charges, if the operators don't block access to VoIP gateways.

Rivals \'could beat BT\' in the next-generation game - ZDNet UK News

Rivals \'could beat BT\' in the next-generation game - ZDNet UK News

The article is related to the ADSLguide one about IP infrastructure and rolling-out all IP networks.

This article goes on about Softswitching and implying it's the best thing since sliced bread. Well it may be.

In a regular telephony switch (like a System X) there's a bunch of intelligence that knows how to handle phone call set-up, call routing etc. and some line cards that actually connect to things and move calls around. It can run out of capacity once the CPU runs out of steam (i.e. it cant handle call processing anymore) or it physically cant handle any more calls through its interfaces.

A softswitch modularises things, you run the call intelligence on some CPU (which is "normal" hardware i.e. Sun servers or equiv) and you splash media gateways around the network. The media gateways interface with legacy PSTN or handle trunking etc.

Why? Because it costs much less to build IP packet networks, than legacy TDM networks (about 40%). Softswitches are also scalable, run out of CPU and you just add some more. Need more PSTN interfaces, add some more media gateways.

However it's not quite as easy as it seems. Softswitches do have lots of functionality, but they are bits of telecoms kit. Say someone wants to launch a VoIP service, just having a Softswitch doesn't give them anything but a platform to build something on. It then gets complicated as web interfaces, provisioning, billing and everything else that telcos take months to actually get working have to be built (if they have the resources to do it in the first place). They may well install a Softswitch just for the cost savings without launching any sexy services on it at all.

ADSLguide: News Archive

ADSLguide: News Archive

Marconi are claiming a rival telco may beat BT at their own game and have an IP only network 1st.

This may well be true, THUS already has a complete IP core, and surely others do too. Unfortunately it's a meaningless comment. BT has 5,600 digital local exchanges in the UK and when they roll-out their 21CN then all of them (or at least anyone that is a real DLE) will be IP enabled. None of the LLU entrants is going into anywhere near this number, most for 3-400, Wanadoo maybe 1000. That's still under a 5th of BT's roll-out.

So until someone has an IP core and unbundles ALL the DLE's, BT still don't have any real competition.