Categories
Engineer internet

IPv4 exhaustion date is Sept 5th 2011

I note that the number of available IPv4 addresses has dropped below the 10% mark. This is displayed on the counter on the right hand column of this blog but it took a link to the Number Resource Org on Facebook to alert me to the fact.

This year should see an intensification of efforts to move to the full support of IPv6. The Sept 5th 2011 date for exhaustion of the IPv4 pool is not very far away now. In reality there will still be stocks of addresses held “in private hands” so that date doesn’t see the unprepared fall off a cliff but it is a clear pressure point.

I should make it publicly known now that I’m planning a party for this date.  Anyone wanting to come along should get their name in early as I anticipate huge demand 🙂 .

Categories
Engineer internet

First 100Gbps commercially available optical network rolled out by Nortel and Verizon

Nortel yesterday announced that Verizon had implemented the first commercially available 100Gbps network on a 893km link between Paris and Frankfurt.

There are several significant (or at least  I think they are interesting) points to be made regarding this milestone.

First of all Nortel is clearly a leader in Optical technology, as it is in a number of its other areas of business. 100Gbps has been discussed at the last few meetings of the London Internet Exchange (LINX) but largely in terms of the fact that 100Gbps equipment has only been achieving 40Gbps, an interim step.

It is a crying shame that the mismanagement of the business during the earlier parts of the decade resulted in the Chapter 11 situation we now see today and the break up of the business. From Timico’s perspective this is at least focussing minds at Nortel and we have seen a significant improvement in responsiveness and keeness to get things done. Good I suppose.

The 10Gbps standard was ratified in 2002 and, doing a quick trawl the first network rollouts seem to be around 2005 –  this is the case at LINX who tend to be up there amongst the leaders. Truth be told it was probably earlier than this.

The 100Gbps standard has not yet been ratified so there are clearly commercial pressures and advantages to running with the technology for a commercial operator to push ahead with it. Historically this has been 4x the cost for 10x the throughput. So it is clear that the cost of bandwidth is going to continue on a downward trend the more people use it, which they are doing.

This is an interesting wave for ISPs and network operators (surfers) to be riding.  We have to be nimble atop the big rollers making sure that we keep our network costs down quickly enough to match the competitive pricing pressures of the market place.

Categories
Engineer voip

VoIP MOS test results are at least as good as PSTN – it's official

The official ITSPA Awards test results make for very interesting reading. All entrants for the Best ITSP, consumer and SMB categories had their services independently tested by Epitiro.

There were 16 entrants for these two categories. On average Epitiro made 400 calls per company and then took over 50,000 technical measurements. Calls were all made over the same broadband connection.

All bar one company tested reached the ITU-T P.862 PESQ MOS Quality rating in excess of 4.0 thus meeting the ITU-T P.800 subjective rating of ‘Excellent’. The one that didn’t met the subjective rating of “Good”.

Consumer VoIP MOS downstream average = 4.3
Consumer VoIP MOS upstream average = 4.25

Business ITSP (SMEs) MOS downstream average = 4.25
Business ITSP (SMEs) MOS upstream average = 4.25

There is no real reason why there should be a difference between consumer and business downstream MOS.

Packet Loss was very minimal. Only three companies experienced any packet loss (minimal – 1.3% was the highest loss)

Call set up times were in general on a par to the PSTN standard of 2.5 seconds and better than mobiles.. The customer would experience no difference.

These are great results and are a serious independent endorsement of VoIP as a mainstream communications technology that can replace traditional PSTN services.

PS MOS = Mean Opinion Score and represents perceived quality of a telephone call.

Categories
Engineer voip

VoIP MOS test results are at least as good as PSTN – it’s official

The official ITSPA Awards test results make for very interesting reading. All entrants for the Best ITSP, consumer and SMB categories had their services independently tested by Epitiro.

There were 16 entrants for these two categories. On average Epitiro made 400 calls per company and then took over 50,000 technical measurements. Calls were all made over the same broadband connection.

All bar one company tested reached the ITU-T P.862 PESQ MOS Quality rating in excess of 4.0 thus meeting the ITU-T P.800 subjective rating of ‘Excellent’. The one that didn’t met the subjective rating of “Good”.

Consumer VoIP MOS downstream average = 4.3
Consumer VoIP MOS upstream average = 4.25

Business ITSP (SMEs) MOS downstream average = 4.25
Business ITSP (SMEs) MOS upstream average = 4.25

There is no real reason why there should be a difference between consumer and business downstream MOS.

Packet Loss was very minimal. Only three companies experienced any packet loss (minimal – 1.3% was the highest loss)

Call set up times were in general on a par to the PSTN standard of 2.5 seconds and better than mobiles.. The customer would experience no difference.

These are great results and are a serious independent endorsement of VoIP as a mainstream communications technology that can replace traditional PSTN services.

PS MOS = Mean Opinion Score and represents perceived quality of a telephone call.

Categories
Engineer internet

LINX67 15th birthday meeting – the internet continues to grow

A milestone for the London Internet Exchange with its 15th birthday meeting today at Goodenough College in London. These are seriously useful meetings of like minded souls from the internet industry.

The internet is run by geeks that work extremely hard. In fact they live and breathe technology on a 24 x 7 x 365 ¼ basis and these LINX meetings bring them together for long days of meetings and presentations but also long evenings of letting their hair down.

LINX meetings are quarterly. Since the last one in the summer the peak traffic on the LINX network has grown by an incredible 25%. If the trend continued that would mean more than doubling in a year. The traditional perceived wisdom for growth in internet usage is 50% pa which is close to what I see on the Timico ADSL network.

LINX’s growth has been fuelled by an increase in membership of around 15% this year though clearly that wouldn’t account for the proportional increase in bandwidth.

For geeks everywhere who believe that size does matters the LINX network currently has over 2.5Terrabit per second capacity. The organisation is clearly providing value for money which is why all these new members are knocking on the door.

A year ago I seem to remember a statistic that said that 65% of the internet was reachable via LINX peering. Now this has grown to 70%. This is good news for end users as it helps to keep prices down.

Interestingly 75% of new members are from outside the UK, including a fair number from Russia. The UK is becoming increasingly an important place on the internet map.

More snippets from LINX67 as I think of them. The meeting is absolutely packed.

Categories
Engineer internet

IPv4 drops below last 400m addresses

I noticed over the weekend that the exhaustion counter for IPv4 dropped below 400m addresses. This is just a stake in the ground. As I write there are 653 days to go.  I’ll revisit this when it drops below 300m.

Categories
datacentre Engineer

Shock Horror – High Performance Laptop Costs $100m

Did you know that a state of the art supercomputer costs $100m?  The price never comes down with time – the speed just goes up. Today’s leading edge box — actually it’s a datacentre full of racks, not a single box — has over 1Petaflops of processing power.

Such is the progress of technology that in three years time this will not even be in the top 500 of supercomputer performances. At that point the maintenance costs also start to ramp up so your average supercomputer owner just bins it and buys another one. It’s what I’d do 🙂

What is also interesting is that today’s supercomputer processor speed flows down to the laptop of 12 years hence. So in 2021 you will get one hell of a bang for your buck. Whether you will need that much power to send emails and operate Word 2021 (or whatever it will be called then) is unlikely.

I guess the power will be usable for improved 3D HD graphics for gaming and TV but I’m not sure what other apps will need it.  Whole brain simulations on a laptop perhaps.  Build it and they will come…  It does point to a huge continued growth in network bandwidth usage.

Of course the laptop won’t cost $100m. I just put that in for effect! My guess is that all laptops will come free with subscription to network services.

Categories
Engineer internet security

Woke up this mornin and nearly got the IMP blues

I woke up this morning (there’s a song there…) to the news from the Daily Telegraph that ” Government announced yesterday it was pressing ahead with privately-held “Big Brother” databases”.  This is the Interception Modernisation Programme that has periodically been in the news this year with general opposition and a subject I have posted on in the past.  It would indeed have given the ISP industry a headache.

I now hear a contrary position from the Guardian which tells me “Legislation to access public’s texts and emails put on hold. Widespread concern about the safety and security of communications data prompts Home Office rethink.”

My understanding is that it is the Guardian that is right on this occasion and that the Telegraph has tapped into the wrong wires. I imagine that the Labour party has enough on its plate in the run up to an election without further alienating the voters.

When they woke up this morning someone got their wires crossed and my head it felt confused, oh yeah. ”  I think that’s what I was trying to say 🙂

Categories
Apps Engineer voip

Strictly Come X Factor needs British Talent – bring on the Media Resource Broker

It’s a fair bet that most punters enthusiastically ringing in to cast their votes on popular game shows don’t think about the network capacity problems they are creating! Typical Joe Public eh?

When someone dials in to one of these shows they make use of Media Servers in the telecommunications network. Typically Media Servers are boxes especially designed for a single purpose. There are a number of such types of server used by Telcos (and I count Timico amongst them) for specialist applications such as the aforementioned IVR based voting system and business services such as conference bridges.

The problem is that the kit used for voting is different to the kit used for conference calls and meetings. This means that expensive bits of kit lie idle in a network for much of the time. Conference bridges are used during working hours and TV Game Shows are on in the evenings (apparently 🙂 ). The surges in network demand prompted by game shows also results in sometimes between 10 and 100 times overprovisioning of capacity compared to the average state of affairs which exacerbates the costliness!

This is all about to change – watch out for the virtual network!

On Friday I met with Chris Boulton of NS Technologies at the Celtic Manor Resort in Newport. Chris used to work in the Office of the CTO at Ubiquity Software, a Terry Matthews SIP company that was bought by Avaya couple of years or so ago and has a blue chip heritage in designing advanced IP communications networks infrastructure.

Chis is currently working on the Media Resource Broker. We are all used to the idea of running different applications on a PC or handheld device, or even games on a gaming box. Telecoms networks though have always been built using dedicated kit. The MRB changes this by allowing telcos to build networks that use standard platforms targetable at different applications according to demand at any given time.

So a box that is used as a conference bridge in the day can be used for voting out contestants in the evening. This will not only save huge amounts of money but also result in a flexible and scalable network architecture that can then quickly be applied to other functions, many of which will not even have been thought of yet.

We talk about doing things in the cloud but it is interesting to see that even infrastructure is moving in this direction.

The whole world is moving towards becoming virtualised. I can even see the day when the typical household will have a box under the stairs next to the gas and electricity meter that will be its network processing resource. This will bring with it huge opportunities in business.

It will also of course heighten our reliance on such resources and the Domesday scenario of when it all goes wrong becomes even more of an issue. But there again even that represents an opportunity…

Categories
dns Engineer internet ipv6

IPv6 DNS progress

IPv6 is on my mind and we have been busy in the background getting ready for the big day. You will have noticed the countdown timer on the bottom right of this blog. Time is marching on.

This last week we have been working on our IPv6 DNS capability. DNS is one of the first steps on the road to IPv6 deployment. It’s all very well having all your routers enabled but if noone can surf to a website, or send emails (to that fridge) then it’s no good is it?

For those of you that know these things:

;; QUESTION SECTION:
;cachev6.dns.as8607.net. IN AAAA

;; ANSWER SECTION:
cachev6.dns.as8607.net. 86400 IN AAAA 2001:40a0:5:53::10

PS if you really need to ask about the fridge then leave a comment or email me at tref at trefor dot net :-). Don’t be embarassed now.  I could just have easily have said light bulb 🙂

Categories
Engineer internet

Hyperconnectivity – modern monitor stands

As we strive to implement best practice for the Network Operations Centre the first major initiative has been kicked off.  Since Fast Ethernet became a commodity it is becoming increasingly cost effective to provide unlimited bandwidth to an engineer’s desk.

At the bleeding edge of technology today at Timico in Newark we made it happen again with the 94 port switched desktop.  Ironically networking equipment is becoming so cost effective that it is now cheaper to double up this kit as monitor stands rather than buy expensive purpose built metalwork that you source from your local office supply company.

Ok I know that the cynical amongst you will decry this as a falsehood.  Of course monitor stands are cheaper but when was the last time you saw a monitor stand doubling up as a network switch.

My only real problem in signing off this minor capital expenditure was the prospect that the two fibre links we have coming into the building will soon be inadequate to support all the extra work that the team says it will be able to get on with now that they have the desktop connectivity.

I am also worried that I won’t be able to give them work fast enough to keep them busy for the whole of their usual 18 hour day. It’s a nice problem to have, as we say in business 🙂

The 96 port switched desktop monitor stand
The 96 port switched desktop monitor stand
Categories
Engineer internet

Netsumo NOC helter skelter

Paid a visit to Netsumo’s Network Operations Centre today with Ben, our Head of Network Operations. We are looking around at best practice in NOC design and it is always a pleasure to spend time with Netsumo CTO Andy Davidson.

Andy is one of life’s genuinely nice guys and very knowledgeable in the Internet Networking space. Netsumo is based at state of the art office facilities at Electric Works in Sheffield and operates over 30 networks on behalf of is customers. The company represents the epitome of best practice for network operations in the UK.

Like all good technical businesses Netsumo looks after its engineers and the working environment is second to none. At centre stage of this is a four storey helter skelter that transports staff to the ground floor in seconds.

Of course we had to try it out – see photos below.

Andy tells us that the helter skelter has two functions. Firstly it is used for rapid deployment of engineers when on an urgent call out. Secondly it is used for rapid deployment of engineers when off to the pub after work. Both are laudible objectives I’m sure most will agree.

netsumo3netsumo1Pics are a little small – I uploaded them to Facebook from my mobile and that is how they turned out!netsumo2

Categories
datacentre Engineer

GigE replaces old ATM infrastructure at Timico Docklands datacentre

I’ve been rolling my sleeves up at our Docklands Datacentres today. Having decommissioned all our old 155 Mbps STM1 pipes and replaced them with 622Mbps STM4’s we are now gearing up to replacing the 622’s with resilient Gigabit Ethernet connectivity to the BT21CN network.

The picture below shows part of the rack containing our first ever 155Mbps connection.  For those interested this was an STM4 partitioned into 4STM1’s.

For those not interested the real point is that this complete rack that was originally pretty much dedicated to hosting our central pipe connectivity to the BT ADSL network can now be replaced by a single port in a 3U chassis. You can get around 13 of these switches in a rack, each with potentially up to 15  GigE connections. In theory that’s up to 195 connections instead of just 4 with 313 x the bandwidth.

That’s progress folks.

STM4 Mux

Old STM4 chassis. Couldn’t get the whole rack in view. This is only half of it.

3U chassis supporting up to 15 GigE connections
3U chassis supporting up to 15 GigE connections

What replaced it!

Categories
Engineer voip

Thus pulls out of VoIP and leaves customers to fend for themselves

Had this link sent to me via the industry grapevine today.  Basically Thus  is closing its “Pipecall” voip service and appears to be leaving customers to their own devices.  This isn’t just a case of people going off and finding an alternative provider.  They need to be able to port their numbers. Typically only BT has porting agreements with absolutely everyone so in this case  Thus’ beleaguered customer base is probably stuck with whatever solution (and pricing) BT might be able to offer them. And it might not be a VoIP solution.

Thus is particularly a special case as the company likely hosted its own numbers on its own interconnect to the BT network.  THUS is not a member of industry trade body ITSPA and does not therefore adhere to the ITSPA Code of Practice in which members endeavour to offer porting to other members.  

The Pipecall user base would almost certainly have had more choice had they been dealing with an ITSPA member company.  Many ITSPA members use wholesale partners to host their numbers and interconnects to what is after all a declining network (ie the PSTN – it doesn’t make sense to invest in “old fashioned” network infrastructure). Thus (not very good pun intended) migration between Communications Providers on a wholesale platform is a lot easier.

As it stands Pipecall customers probably now need help with their comms. If anyone wants to get in touch for advice feel free to contact me  directly. I’ve pasted below the content from the Thus/Demon website on this subject for ease of reading. Oh and by the way Timico is just investing in a major VoIP platform infrastructure upgrade.  Watch this space.  We are in it for the long run.

Why is my account being cancelled?
THUS has made the decision to close the platform which carries the service you are currently using.

Why are you putting an end to the service?
The platform has reached the end of its serviceable life. It is not possible for THUS to effectively maintain it and deliver acceptable levels of reliability, performance and security and as such, the decision has been taken to close it.

What are you putting in place or what are you offering to do with our numbers?
There will not be a directly comparable Voice Over Broadband service available from THUS. See below on how to port your numbers to an alternative provider.

What is being done with my details that are held in your database / system?
These will be deleted.

How do I port my numbers?
Customers wishing to port their numbers should contact BT. If they select an appropriate product from them, BT will port the numbers over themselves when the service goes live.

What will happen to the personalised IVR on my service?
As the service is closed all IVR functionality will be lost. Customers will need to replicate what they had with their preferred new service provider.

What will happen to my stored faxes etc on the control panel?
This functionality will be lost with the service termination.

Will my hardware work with any provider?
We can’t guarantee that it will. However the Voice over Broadband service was built using industry standards, so should your new service also be SIP based, your hardware should still function.

What if I need to collect previous CDR data or invoices post the closure date?
These will be available for 3 months afterwards.

When will the system close?
The system is due to close at the end of September.

I’ve heard about the cease of service but I haven’t received my letter of notice?
The letters were all sent out to our customers registered addresses at the end of August.

While the numbers are being ported, will that cause any downtime to my service?
No. Should your new provider follow the correct procedure there should be no downtime.

Will support still be available after the closure date?
As the service will no longer be operational support will not be available.

I am still within my 12 month contract, will that cease also?
Yes

I am in credit on my account, will that be paid to me in full?
Yes – Please contact Customer Services by email via [email protected] or via telephone on 0845 009 0080 (Contactable between 9am-5:30pm Mon-Fri) so that we can arrange for any outstanding payments to be made to you.

How long does it take to port a number?
You will need to speak to your new provider to find out the timescales.
The minimum lead time is 5 working days but your new provider may take longer.

Categories
Engineer internet media

Video streaming is going to be a mix of technical and commercial problems

Interesting talk  at Linx66 by Giles Heron, Principal Network Architect for 21CN at BT. I’m not going into it all but towards the end he discussed video streaming on the network and ways that this might be made more efficient.

Whilst ISPs and network operators take on the challenge of managing the growth in video streaming I think there is just as big an issue coming along in respect of the commercial and regulatory aspects.

The time is not far off, I believe, where people will stop buying purpose built boxes and rely purely on their internet connection for their TV watching. My kids already do it. The advent of FTTC is going to accelerate this. If you watch TV over the internet the rule is in the UK that you are supposed to have a TV license. This is not currently policed as far as I can see and it is difficult to do so. Last night someone in the family was playing a game on our TV so I watched Usain Bolt do his stuff on my laptop.

It will not be long before the BBC will start seeing its license revenue take start to decline (maybe it already is) which is going to prompt a very high profile crisis. We are going to start getting the same debate re TV watching as we have been having with music piracy – who is going to police it?

The alternative is that the BBC will have to move to a pay as you go model for iPlayer watching.

Well done to Usain by the way if you are reading this. Very good 🙂

Categories
datacentre Engineer peering

Interexion talk on green datacentres at Linx66

Some interesting talks at the today’s Linx66 sessions at Goodenough College in London. Lex Coors, VP of international datacentre operator Interexion discussed the green datacentre. One of the slides that caught my attention related to best practice in how end users can keep their power consumption to a minimum.

Most of these are pretty obvious but worth reproducing here with the percentages being the potential efficiency gain:

eliminate comatose servers 10 – 25%
virtualise 25 – 30%
upgrade older equipment 10 – 20%
reduce demand for older equipment 10 – 20%
introduce greener more efficient servers 10 – 20%

If you add that lot up you potentially get more than 100% but it does give people a feel for where their efficiency savings and therefore cost savings can be made.

The original source was McKinsey and the Uptime Institute (2008).

Categories
Cloud Engineer internet

Virtual Machines

I’m doing some research into Virtual Machines, storage trends and cloud computing. I’d be interested in hearing from people who use these services, thinking about using them or from software vendors offering solutions in this space.

I’m particularly interested in who is using such products in the UK. Timico already offers these types of services but I’m looking at expanding this area of activity and would welcome input with product definition and market segmentation.

Anyone wanting to contact me can either just leave a comment on this post or email me (tref at timico.co.uk) and I will happily call them back.  Thanks in advance.

Categories
Engineer internet social networking

Twitter downed by ddos attack

I have to apologise to Dave Ward who manages the firewall at Timico HQ in Newark.  I complained to him that he was blocking me from accessing Twitter and he scurried off to check having denied it all. (it’s not a criminal offence in my book anyway).

I just read that Twitter was this afternoon hit by a distributed denial of service (DDOS) attack which took out the service for a few hours from around 2pm.  Sorry Dave.  Whenever I publish a blog post Wordpress automatically sends a tweet on the subject which in turn updates my Facebook status. What would we do without Twitter eh?

Categories
Engineer internet spam

Anti spam best practice

You may have noted the spam theme of my posts this week.  This is because we are in the process of upgrading our anti spam capabilities. The management of spam is a hugely complex process and involves many factors contributing to a scorecard against which an email is rated.

 

There is a general set of principles that the industry could apply that would make it a lot harder for spammers. Unfortunately many ISPs seem to the fairly lenient with their customers about how they set up their email services and are prepared to accept mail from poorly configured mail servers.

 

For example most spam comes from compromised Windows computers at residential or business premises.  When a host connects, ie when a mail is being set up for sending, it should perform a HELO with it’s fully qualified domain name (FQDN) as specified in RFCs (industry standards or standards in waiting).  The sender sometimes lies and presents a fake or incorrect HELO string, which can be used to judge the validity of the sending server. The string given at HELO time should have forward and reverse DNS that matches. 

 

Additionally, the reverse DNS of the sending host could be considered.  If there is no reverse DNS, it’s very unlikely that the mail is legitimate, and should be rejected.  If the reverse DNS makes it clear that the sending host is within a DSL pool, ie at the user premises at the end of an ADSL line rather than an ISP’s mail server, this could also be taken into consideration when it comes to scoring.

 

A genuine Reverse DNS might look like mail.timico.net whereas a corresponding ADSL based DNS, (and therefore likely to be the source of spam), would be xxx.xxx.xxx.xxx.adsl.timico.net where the x’s represent the ip address.

 

Another technique in the fight against spam is to rate limit emails from users. In other words to apply a policy controlling a maximum number of emails an individual can send in a day.  A rate limit for a residential user might be 200 mails a day for example.  It is unlikely that the residential user will send more than 10 or 20 mails in a day.  A compromised machine may, however, send thousands in the same time period. The rate limit would prevent this. 

 

Customers with a genuine need to send more emails than the limit can easily be accommodated.  The limit is there to protect the user rather than to stop them sending emails. The spam being sent would normally be caught here anyway but this technique does at least minimize the load on spam filters.

 

The factors taken into consideration in spam scoring systems are not normally made public domain because to do so would just help spammers.

Categories
Engineer spam

More spam

As an update from yesterday’s post the botnet attacks are getting even more frequent. I’ll post some notes tomorrow on best practice for ISPs contending with spam. You can see the increase in frequency and intensity over yesterday. Some crook somewhere has obviously decided on a new “marketing campaign”.

mail-darktues

Categories
Engineer internet spam

Spam attacks

I sometimes sit and watch SPAM attacks coming in on our mail servers. ISPs are constantly having to ward off spam. It is like being in a cyber war. What is mind boggling is that sheer volume. The chart below shows unwanted mail below the x axis and legitimate stuff above.

You can barely make out the legitimate mail because in the scale it is dwarfed by the spam. It is also interesting to observe that the attacks come in waves.  You can see the major incoming waves on Sunday evening followed by periodic smaller attacks which appear to be the work of botnets.

This spam is of course not passed on to our customers who pay us for a premium service. You can zoom in by clicking a couple of times on the picture.

mail-dark1

Categories
datacentre Engineer

It’s all about wiring

Following my post on our fibre installation earlier in June The build of our new datacentre module in Newark continues.

Datacentres, whilst giving the appearance of being high tech,  are all about wiring and plumbing.  So I’m getting in the cable monkeys and plumbers.

Couple of photos below give you a feel for part of the process. Underfloor power connections to each rack space and a coil of fibre that might look innocuous but will carry the lifeblood of the datacentre, ie the data itself.

It makes me think of the pony express, or the old stage post mail system and how things have changed. I’m getting romantic in my old age.

cabling

 

 

 

 

 

 

 

 

 

fibre

 

Of course it will be tidied up a bit before we open for business.

Categories
datacentre Engineer

It's all about wiring

Following my post on our fibre installation earlier in June The build of our new datacentre module in Newark continues.

Datacentres, whilst giving the appearance of being high tech,  are all about wiring and plumbing.  So I’m getting in the cable monkeys and plumbers.

Couple of photos below give you a feel for part of the process. Underfloor power connections to each rack space and a coil of fibre that might look innocuous but will carry the lifeblood of the datacentre, ie the data itself.

It makes me think of the pony express, or the old stage post mail system and how things have changed. I’m getting romantic in my old age.

cabling

 

 

 

 

 

 

 

 

 

fibre

 

Of course it will be tidied up a bit before we open for business.

Categories
Engineer internet spam

Spam 2.0

I don’t know about you but I have started getting spam through Facebook. So far it isn’t the classical type of spam selling viagra etc.  I have however been getting friend requests from attractive young ladies with exotic names. 

I also seem to be inundated with notifications of rubbish that I have no interest in checking out.

I was discussing this with Dave Ward, one of our Tech Consultants, who mentioned that Fortinet have now brought out Spam2.0 filters for their firewalls.  Social Networking sites have started having their vulnerabilities exploited. 

People are getting spammed with direct messages, apparently from friends.  Facebook chat, for example, is one way used to insert worms onto someone’s PC and thence onto your network.

Fortinet has an application that allows companies to let employees access Facebook whilst blocking access to applications such as chat known to be vunerable.  Screenshot below. You might need to click a couple of times on the picture to get it to a viewable size/quality.  Also check out the recent Wikipedia article on Social Networking Spam.

spam20

PS Don’t get me wrong here.  I am a happily married man and whilst I’m sure I quite like being chatted up by nice young ladies one has to ask why complete strangers, whose interests seem to be dating and meeting members of the opposite sex, would want to approach me…

Categories
Engineer internet peering

#LINX65 notes on traffic growth

LINX CTO Mike Hughes told us that traffic across the LINX network has reached 460Gbps.  Thats up 40Gbps in the last quarter and around 60% up year on year. 

The growth is partly due to new members and whilst this isn’t a definitive statement on the growth rate of the internet it does match quite well with other sources.

In the last year the growth pas particularly come through an increase in the number of 10Gbps connections, up now to 175 ports.  The LINX team is already preparing for 100Gbps technology, the standard for which is due to be ratified next year.

linx65

The chart clearly shows the growth of the number of 10 Gig ports over the last few years.  It’s a bit grainy but clear enough. The network capacity is now north of 2Terrabits per second over 650 ports.

Categories
datacentre Engineer internet peering

LINX 65 and Telecity

First day of LINX65 produced the usual interesting mix of talks. Today included IPv6 and VoIP QoS.

The sponsor’s talk at the end was given by Rob Coupland, COO of datacentre operator, Telecity. In Europe Telecity operates in London, Paris, Amsterdam, Stockholm, Milan and Frankfurt. A good footprint to have.

What was interesting was the statistic he floated that the company is doubling its datacentre power capacity over the next couple of years.

I counted 26.5MW in total! They plan to sell this over the next 3 – 4 years. This is a big bet that they appear to be confident of placing based on the uptake that they are already seeing. One of the big drivers they are (unsurprisingly) seeing is content provision.

I’m not making any comment re the effect on Global Warming here seeing as we at Timico are also in the business. I guess at the scale that we are talking about though cooling efficiencies will make a huge difference.

Categories
Engineer internet voip

IETF 74 and SIP

10 years ago this month saw the publication of RFC2543 which was the first proposed version of the SIP standard that is now used an almost all internet telephony services.

This is being celebrated this week at IETF74.  The Internet Engineerng Task Force is the body that maintains standards  for internet related technologies.

SIP was initially championed by a small number of people that included it’s inventor Henning Schulzrinne, Jonathan Rosenberg, Jiri Kuthan, Henry Sinnreich et al.

At around that time I was being asked by my then employer Mitel to set up a product line that was based on open standards.  There was quite a choice to chose from.

MGCP was adopted by the cable community in the USA and by a number of ITSPs.  However the problem with MGCP was that it had a relatively small feature set which meant that service providers had to develop their own extensions to provide saleable services.  Thie meant that MGCP quickly became non standard as any venbdor would have to support multiple flavours of the protocol.

Then there was SGCP, or skinny. This was a Cisco proprietary protocol.  Whilst potentially this had initially the largest market opportunity it did tie you into Cisco.

So SIP was the obvious one but it took a long time for the market to appear, particularly as the turmoil around 9/11 hit the dot com intustry.

I spent 4 years service on the board of the SIP Forum from around this time.  Being around during the early days of a technological revolution was exciting and I am fortunate enough, with Timico, to be able to continue the ride.

Check out the Facebook event surrounding this anniversay here.

Categories
Engineer internet

The Internet Protocol Journal

Geeks are a great breed and I am pleased to say our NetOps team is 100% staffed by them.  They are a real asset to our business.

If you ever want to test whether someone is a geek or not you have to see what reading material they take to the “smallest room” . 

At the Timico NOC there are copies of Cisco’s “The Internet Protocol Journal” left there to ensure optimum use of time.

The journal also provides the facility with an N+N level of resiliency making it a Tier 4 toilet.  I’m pleased to say we have never had to resort to a failover.  Read on!

Categories
Engineer internet olympics

Surreal Sewer Story

The things I have to do for the business. Today it was almost like being in a James Bond movie. I turned up at 10am prompt at the Thames Water Utilities Depot in Bow in East London – just around the corner from the site of the Olympic stadium which incidentally seems to be shooting up. I was met by Calypso Harland, bubbly marketing manager of Geo, the Alchemy owned business that sprang out of 186k’s UK fibre optic network.

I had been invited to inspect Geo’s London fibre ring which, yes you’ve guessed it, runs through the sewers. Donning layers of protective clothing I disappeared into a subterranean world for a once in a lifetime experience. Believe you me I doubt that anyone wants to do it twice.

The 140 year old Victorian sewers under London are I suppose of cursory interest. The message actually was that Geo can provide very cost effective and very secure fibre backbone connectivity because they use the sewers and don’t therefore have to dig up the roads.

What is interesting is the huge capacity there is down there. Geo run 4 ducts, three of which are currently empty. A duct can hold two fibre cables each with 288 strands of fibre. Each fibre can, using current DWDM technology, run 160 10Gbps channels (x 4 if you use phase modulation per wavelength).

So by my sums the Geo London ring should be able to operate at nearly 15 Petabits per second or roughly a billion times faster than the latest and greatest 21CN ADSL2+ connection. That’s a heck of a lot of capacity.

And for anyone that wants to know no, my thigh length wellies did not leak and no there were no floaters in sight – something to do with a high fibre diet apparently! 🙂

These sewers run on the surface (covered up obviously) right through the Olympic site. Come 2012 they will likely be crawling with security guards!

Check out the Geo video here.  Photos below:

Categories
broadband Engineer internet

Broadband Growth Results in Network Capacity Increases

Been a busy week again. Where do they all go? This week was a milestone for Trefor Davies in that I got my first 622Mbps BT central pipe. It’s not mine really of course, it is Timico’s but it feels like mine – toys for the boys 🙂 .

Not everyone will understand the significance. When an Internet Service Provider starts life they begin with a small connection to the BT (Wholesale) ADSL network. They use this connection, known as a “central pipe”,  to carry broadband web traffic from their ADSL customers to the internet, email server, co-located equipment etc.

The ISP itself usually provides the applications (ie email) and contracts with a variety of network providers to offer resilient access to the internet.

As the number of broadband customers grows more BT central capacity needs to be added and this is normally done in affordable “chunks” rather than going straight for a big connection that would almost certainly be uneconomic for low levels of ADSL customers.

Unfortunately for the small ISP the bigger the connection the better the level of service experience  that can be offered. This is particularly the case if an ISP has many small (34Mbps) pipes.

So this week Timico moved into the bigger league with a 622Mbps pipe and cancelled 7 smaller pipes. What is more our second 622Mbps pipe is due for delivery in April giving us a huge headroom in capacity/expansion capability. This is in additional to the resilient 1Gbps links we have for 21CN.

Ironically although supposed to provide 622Mbps these pipes in reality provide more bandwidth than advertised. This compares with a 155Mbps pipe which typically only provided 120Mbps capacity in two separate 60Mbps halves and a 34Mbps pipe which only provided 25Mbps capacity.

Exciting eh? ! 🙂