AMS-IX breaks 600 Gbps

The previous post about AMS-IX embedded the traffic graph instead of displaying a snapshot. At least I noticed that AMS-IX now pushes over 600 Gbps [peak] of traffic through the exchange!


But what's this? Number 2 exchange DE-CIX claims a 650+ Gbps peak:


Does that mean the Deutscher Commercial Internet Exchange is now #1?
Comparison of the mean throughputs (AMS-IX: 375 Gbps, DE-CIX: 299 Gbps), daily peaks (AMS-IX: 560 Gbps, DE-CIX: 460 Gbps), and number of interconnected members (AMS-IX: 311, DE-CIX: 264) shows AMS-IX still firmly in the lead. But DE-CIX is forging ahead - AMS-IX's lead doesn't seem as clear as it once was.

AMS-IX breaks 500 Gbps!

[from the AMS-IX status page]

The number 1 internet exchange in the world by traffic volume - AMS-IX in Amsterdam - has reached 500 Gbps!



Notice the well-known flat summer trend. It's interesting to project what the max will reach by December: 550-600 Gbps?

Amazon's new CDN

Amazon's new CDN is cheered on by their CTO Werner Vogels. The primary take-away from the post:

Using a global network of edge locations this new service can deliver popular data stored in Amazon S3 to customers around the globe through local access.

Now, Amazon is only currently peering visibly with 4 other networks: Qwest, NTT, Tiscali, and Level3, which all seem to be transit providers.

Now I guess their edge routers don't absolutely need to sit in peered networks, but I'm sure it'll make the host networks happier [otherwise those host networks will be bearing the cost of reverse-proxying the traffic through their own transit connections].

The PeeringDB page for Amazon isn't that much more helpful - no public switching presence noted, but there is some promise shown in the following locations:

• Equinix Ashburn
• PAIX Palo Alto [incl. PacketExchange I guess]
• The Westin Building [Yay SIX! And PAIX Seattle too, I suppose]

Another development to watch closely. Amazon: where is your non-USA PoP presence? LINX and AMS-IX beckon...

TENET peers with Google

TENET's UK frontend Ubuntunet [wiki] is now peering with Google.

This is only in the UK for now, presumably they'll peer locally when Google gets their act together. Let's hope Google isn't waiting for the Seacom cable to land too.

Google local delays

So now SM of Google SA is promising better SA-local access to Google services by the end of 2008. Only a year after they promised it the last time:

[via mybroadband.co.za]

Masie said that by the end of 2008 services such as Google Search, Google Maps and Youtube would work a lot faster, better and quicker in South Africa.

Google SA have been quite coy in announcing anything about their local datacenter plans. Is this to avoid tipping their hand? Perhaps they'll extend their GGC 'CDN' to South Africa. This might be the best way to dip their toe in the waters and gauge local usage and interest, especially once mid-2009 rolls around and Seacom's better pricing comes online.

Don't care about nanowires? That's good.

[via DailyTech]

So UPenn researchers have discovered a good way of storing ternary digits.

Great news for Cisco, Juniper, et al. You may wonder why, and the answer is quite simple: TCAMs. Basically this could help speed up address mangling, e.g. for routing tables.

Facebook making money via Google

Does anyone else see a Google Custom Search box beneath the applications list?

Facebook could be raking in some serious money from that.

Surprise surprise

So it turns out that some people have been using the looming spectre of the 2010 Soccer World Cup to help talk up their projects.

Infraco - I'm talking about you!

It's not that surprising, really. What is interesting is the talk of SAT-3 (amusingly and unintentionally further abbreviated as at-3) being upgraded to 320 Gbps. This will at least help Telkom with some extra bandwidth - 120 Gbps was pretty paltry at introduction and now looks downright anemic. Does Telkom dominate the SAT-3 consortium so much that it can just force the upgrade?

Isn't it procrastic?

The irony escaped me yesterday; I guess procrastination when it comes to blog posts isn't good...

Procrastination is good

[thanks HPCwire]

Says [Chief IBM Roadrunner engineer Don] Grice: "A job that would take you about a week to run on Roadrunner would have taken you 20 years to run on a machine just 10 years ago."

Gro'in' pains

Enough 80's references, let's skip to the meat of this post (so to speak). This is from a recent traceroute; DataHop might want to rename one of their nodes to accomodate some expansion in their traffic:

11 196.32.209.25 (196.32.209.25) 165.620 ms 160.809 ms
12 ge2-20-0-cr0.tch.uk.as6908.net (78.41.155.161) 236.578 ms 198.183 ms
13 te-4-2.r00.londen05.uk.bb.gin.ntt.net (83.231.146.69) 151.840 ms 153.098 ms

UCT's smallest office

This is one of the candidates, for sure.
And also a candidate for "Least Paperless Office of the Year"

[and I'm tempted to say that it was so small I couldn't fit inside to take pictures, but I was actually too lazy to try]

[office extends from the pillar on the left of the frame to the right of the frame - i.e. 2 window frames]

[the third window frame visible on the right is someone else's office!]

This is in the Leslie Commerce building. At least this is a hard worker (all those papers!), so no comments about the tightwads there, please...

The New Stephenson opus

The new Neal Stephenson book should be out within 6 months!

My guess is it's going to be the most exercises some geeks will get for a while - 928 pages!

Facebook CDN

I don't know how this slipped through the cracks, but I've just noticed that Facebook has started hosting some content on static.ak.fbcdn.net [fbcdn presumably = facebook content distribution network]

Is Facebook planning on using multiple CDNs?

Both static.ak.fbcdn.net and static.ak.facebook.com seem to be hosted by Akamai, so for us "lucky" .ac.za we get routed through IS. This is actually good news, as we have a very high-speed connection to IS from TENET. Unfortunately, most institutions connect slowly into TENET (~30 Mbps for UCT!) and implement some shaping on popular websites. So we don't get most of the benefit.

Science bandwidth requirements estimate

A recent e-mail from a Meraka engineer working on SANReN:

We know eVLBI would require a 1Gb all the way to Amsterdam, current projections for full SKA is 100GB international capacity, the MeerKAT project will probably require 10GB of bandwidth.

So something like the 'bandwidth glut' of Infraco would definitely be required.

It seems like the only people talking about having too many cables are those with a vested interest in having only a few [competing] cables.

Extra transit for TENET

Viewing the monitoring graphs for TENET's London node, a new transit circuit through DataHop [web?]is apparent:


Peak speeds are about 50 Mbps. Combined with the Telia peak of ~100 Mbps, this means TENET currently has ~150 Mbps peak transit.

Note the very small outbound. Mmmm, eyeballs.

Unscheduled scheduled outages

A recent posting to the TENET ren-news list:

All Neotel MPLS GEN3 sites experienced a number of service failures
early yesterday morning. It appears that these were caused by Neotel
maintenance, of which TENET was not informed in advance.

even given

The Service Level Agreement provides that: "...The Providers will give TENET four days' advance notice of planned maintenance."

Let's hope Neotel doesn't become Neotelkom [the new Telkom?] :)

ICTS vs users

For some reason there is an article praising UCT ICTS's War on Terror...er...Bandwidth:

[snip lots]
ICTS has implemented Packeteer Packetshaper
[snip lots more]

AFAIK this actually happened ages ago. So there's no real reason to trot this out now unless people are griping that GEN3 isn't doing much for them.

And there's no mention of what will happen once:

• UCT transitions to SANReN and has a 10 Gbps link to the local node, and
• TENET moves to the 10 Gbps overseas pipe due mid-2009

I don't see the Packeteer appliance scaling to multi-Gbps speeds without serious extra $$$ invested. And would it be worth doing so?

Consider this paper on web search clickstreams http://doi.acm.org/10.1145/1177080.1177110]:

The MWN [Munich Scientific Network] provides a 10Gbit/s singly-homed Internet connection to roughly 50,000 hosts at two major universities along with additional institutes; link utilization is typically only around 200–500 Mbit/s in each direction.

Now this might have been back in 2006; it doesn't really matter if the traffic demand has doubled since then. Also, since major local and international peering will then be in effect, actual transit pressure should decrease.

C'mon ICTS - plan ahead now!

Happy birthday to me

Thank you to everyone that wished me a Happy Birthday (and there were a lot of you), far-flung* or not.

I did have a great day, and I got loads of really cool (and some really cool and unusual) presents.

Based on my experiences with this birthday, I can definitely recommend having a birthday to anyone. I myself plan to have many more.



* Whomever said that technology depersonalises our social interaction was a 'nana.

More Firefox and ISA

The advice in the previous post about Firefox interacting badly with Microsoft ISA seems to be outdated.

The correct advice now:

1. Open (i.e. type) about:config in the location bar
2. Search for network.negotiate-auth.allow-proxies in the search bar
3. Double-click the line to set the value to false

You should now be able to enter your username/password combo once, have it remembered, and not be pestered with millions of pop-ups.

Flatlining

With all of the negative feeling towards TENET and GEN3 because of the recent outages, I thought it might be quite amusing to show the reverse:



CPUT's main Bellville campus is obviously in need of more, More, MORE! bandwidth. Roll on, SANREN.

150 GB DSL caps?

From the latest DSL Prime:

My follow-up article will be unreliable sources, which I'm factchecking more carefully. For now, a few types to distrust:

• The FCC or anyone else who takes seriously a 200K definition of broadband.
• Anyone who defends a bandwidth cap below about 150 gigabytes per month, except where bandwidth costs are unusually high or the service selling for under about $20. (India, South Africa, possibly the U.K.). Ask mo [sic] for the numbers if you don't get it.

I guess for South Africa, he's specifically talking about high bandwidth costs, 'cos I don't know of any ISPs selling DSL for $20/R160 here.

GEN3 failure

And so soon after the start of the service?

Hmmm, what does the SLA say about this? 99% uptime implies 1% downtime.

365 days x 24 hours x 1% = 87.60 hours down per year.

So they have loads of downtime left - bah humbug.

A quick estimation from the ML Sultan campus of DUT:


The interruption starts around 19h30 on the 5th of March, and lasts until about 09h30 on the 6th of March. That's 14 hours total. Ouch.

So we can look forward to another 73.6 hours of downtime. That's another 3 days left!

GEN3 draft SLA

GEN3's draft SLA is available (and claims to have been posted on 27th February).
Interesting to see specific figures. I'm intrigued to see that 260 ms RTT 'average'.

5. Availability

The Providers commit to the following levels of service element availability:

Services provided by Neotel:99%

Services provided by Internet Solutions:

International services: 99.8%

National services: 99.95%

6. Response times (“RTT”)

The Providers commit to the following network RTT averages:

Services provided by Neotel: 150 ms

Services provided by Internet Solutions:

International services: 260 ms

National services: 35 ms (services provided out of Rosebank)

50 ms (services provided out of Cape Town)

7. Packet loss

The providers commit to confining packet loss to the following levels:

Services provided by Neotel: 1%

Services provided by Internet Solutions:

International services: 1.5%

National services: 2%

GEN2 to GEN3

Is UCT currently operating on both GEN2 and GEN3?

Compare GEN2 stats vs GEN3 stats:

The big view seems to indicate traffic starting sometime in Week 07, about Tuesday:


The uncommissioned orders page shows UCT's link to become active on 1st March, 2008. So maybe it's all testing traffic.

It's interesting to see the TENET-JINX graphs too:
Note that we're not seeing massive usage through JINX. This basically boils down to two costs as limiting factors:
1) The equivalent-line-charge
2) The ISPA category charge (small, medium, large)

Now TENET is an honorary ISPA member, so presumably doesn't have to pay the small [by comparison] R6100 pm. But there's an incredible step function in the Equivalent Line Charge going from R0 per month to R22k per month as you transgress 2 Mbps. So can we assume private peering in/near the JINX facility?

Google's local servers

I don't know how this slipped by me, but it's finally appeared in my Google Alerts:

Stafford Masie of Google SA at a MyADSL conference in November 2007 [thanks ITWeb]

“We are going to establish a point of presence in this country, which means the international bandwidth problem is solved. There will be a big announcement in a month,” Masie said at the time.

Google do host a lot of bandwidth-intensive services, and having these served locally will help a lot with latency (for the cached stuff, of course) and cheaper bandwidth, especially for consumers.

Those local-only ADSL accounts are starting to look great. Telkom's partial-compliance with the ICASA rulings on local bandwidth will help too. Perhaps using Stefano's DSL split-routing setup will become more commonplace.

Of course, a month from November 2007 was December 2007 or, charitably, January 2008. Still no word though. Google - what news?

---

Let's see, the more bandwidth-intensive Google services are:

• GMail
• Picasa and PicasaWeb and Image Search
  
• Docs (watch out Microsoft - your anti-piracy drive may backfire)
• Maps
• Google Video (and does this mean YouTube will be served locally too? What about Limelight?)
• Google Earth (mmmmmm)
• full list of all services
  

Night-time power usage, redux

So the newspapers [Business Report, at least] have finally caught up.

Power rationing could result in demand for electricity falling by between 10 percent and 20 percent if successful and could dent Eskom's revenue, presenting further difficulty in funding the company's capital expansion.

You read it here first, folks! So I've been ahead of the game on that call :) I look forward to breaking many more stories for y'all.

Roll on Electric-gates!

AMS-IX tops 400 Gbps

AMS-IX finally tops 400 Gbps (this happened a few days back, but it's taken me a while to get in front of a PC):


Now the daily graph does only show a 395 Gbps maximum, so this could be [partially] an aliasing issue, but I think it's safe to say that the milestone is passed.

Eskom vs corporate consumers

Why are people so upset with office blocks leaving lights on the whole night?

It's not like Eskom can save all of that power for use during the day; Palmiet can only generate a modest amount of power: the 2 turbines are only 200 MW each.

And who else is using lots of power at night? Not household consumers. So there is more than enough to go around at that time.

And who else is using lots of power during the day? Not household consumers. Many of them are at work / school / out-of-the-house.

So the primary contention-period is during early-morning periods (with hot water cylinders, kettles, and microwave ovens going mad) or evening periods (with hot water cylinders, stoves/microwave ovens, televisions*, and heaters/airconditioning being the primary culprits). And many people are early into their offices (I used to get in before 6am myself sometimes) and leave late (oh those productive captains-of-industry). So I don't think many people would complain about power usage in offices during those periods.

So what's all the complaining really about?
"Those people have lights while we don't. They don't need the lights and we do. They suck."

Back in reality, we should encourage conspicuous consumption by large corporates. I would love them to be sucking down as much juice as possible during the quiet period, as this means more demand - AND MORE PROFIT - for Eskom. Would you rather have Eskom build all of those new power-plants purely from your tax money, or from the profit they have earned off other people's excess consumption? No one else needs that power in the middle of the night.


* - Have you noticed that the power-consumption-warnings on TV say to turn off all nonessential appliances except your TV? Hmmm...

Chocolate Mousse

CHOCOLATE MOUSSE

Ingredients:
300g dark chocolate
4 eggs, separated
25ml Kahlua
250ml cream, whipped
5ml gelatine
40ml water

Method:
Break chocolate and melt.
Sprinkle gelatine over 40ml water and allow to sponge (dissolve.)
Whisk egg yolks until fluffy and add melted chocolate and liqueur.
Add dissolved gelatine and mix well.
Whip egg whites until stiff and fold into chocolate mixture.
Fold in whipped cream.
Put in glass and chill until set.
Decorate.

~ ENJOY ~