Now, thanks to the generosity of my friends at IPINX, this blog is hosted in Midrand, Gauteng and it has native IPv6. I would guess this is probably one of the few web servers in South Africa with native v6 outside of AS2018.

I took the opportunity to upgrade to the latest WordPress. If you spot anything thats broken, please let me know.

Graham, among many others, has done a good job of describing how Apple has crippled the iPhone in this regard by locking down the platform and imposing a broken app store model to “publish” software that runs on it. Instead of re-hashing the ecosystem issues surrounding mobile phone platforms, I have a technical wish list:

1. Hardware and software feature parity with best smart phones out there:
   • Quality touch interface, high resolution display, good battery life, all the right radio hardware (GPS, GSM, Wifi, UMTS), sufficient storage and CPU power
   • Software to support standard phone use-cases out of the box: turn-by-turn navigation, tethering, web browsing, full multimedia support, camera functions, etc.
2. The same Open Source UNIX-like operating system I use on other platforms: this happens to be Linux
3. A familiar, quality package system like Debian’s.
4. The ability to cross-compile and run any application I could use on my PC (this implies X11 support).
5. Support for standard peripherals (USB host/otg).

Neither the iPhone, nor any Android-based phone meet these criteria.

The Nokia N900 comes very close, but it’s still missing software support for some standard smart phone use cases. Although I’m a bit disappointed about the Maemo -> Meego move throwing dpkg out in favor of RPM, I still have hope for the Nokia, and I’m interested to see them finish the software (Hi Edwin).

Finally, I want hardware support for an external display. That might sound weird, but if a phone is really a general purpose portable computing device, then you should be able to “dock” it at your desk, like one does a laptop. Maybe I don’t “get” smart phones, but I prefer the view that the world is wrong.

Anyone with me on this? I would appreciate some comments.

Update: The Nokia N9 prototype running Meego seems promising. I guess we’ll see when it’s launched.

The wikipedia article describes OAN in more detail, and refers to a number of examples where this model has been implemented in broadband access networks.

But if the service provider doesn’t pay for the infrastructure, who does? Governments have already done this once, discovered they’re no good at it and consequently privatized the state owned telcos. These telcos are moving in the opposite direction of open access, instead trying to figure out ways out of the dumb-pipe trap through vertical integration, value added services, and so on.

An idea called homes with tails proposes to include the last mile as part of your property when you build or buy a house. This is smart, because it opens up a brand new (and also bandwidth/throughput/services independent) way to amortize the cost of the broadband access network.

Yesterday, Joe Botha blogged about what he imagines a network architecture/billing model for broadband OAN could be. What he doesn’t mention is that this idea is not new to South Africa. This prompted me to dump my thoughts on the matter (many of which I’ve been ruminating on for some time) into this post:

A few years ago, Neology (who I’ve previously blogged about) did an OAN pilot project in Hatfield for the Tswhane (Pretoria) municipality. If I have the story straight, this was a Wifi based mesh network that had an http captive portal that let you choose between 2 or 3 different ISPs, and it worked. This initiative subsequently stalled (apparently due to council politics) but my point is that OAN broadband been demonstrated in the local context.

If the access network infrastructure was not owned by government, but by the end users instead, perhaps the outcome would have been different. So let’s push this idea further. The next best thing would be to avoid relying on local government for wayleaves/right of way/servitudes. This might be possible using wireless local loop, but I’m just going to ignore that option because a bunch of home owners have little chance competing against multi-billion Rand tel/cellcos for exclusive access to radio spectrum (and the ISM-band WISP approach has some drawbacks).

Now it only takes a short drive around your city in South Africa to notice the transformation from traditional suburban townships to gated communities/housing complexes/residential estates. Herein lies the opportunity to do exactly what I’ve described in the paragraph above. I’ve head some entrepreneurs and local ISPs (who?) have begun to wire up these estates internally and provide centralized Internet break-out. I don’t know if this is done on an open access basis, but again, my point is that this piece of the puzzle exists.

Finally, Dark Fiber Africa have recently begun doing something similar in office parks, and as with their metro dark fiber networks, their model is open access.

So where to from here?

Vodacom, MTN and Cell C have all invested in fiber-based back-haul to their cellular base stations (ever wondered why the roads are being dug up and trenches filled with brightly coloured plastic ducts, deep in residential areas?). These base stations now represent an extensive “curb” (in the network sense), and since in many cases towers are shared by two, or all three operators, I think base stations make an ideal location for OAN broadband “meet me” points.

I haven’t asked, but I suspect the cellcos aren’t thinking this way (yet). If they venture into fixed line residential broadband, they will probably first attempt PON-based overbuild in affluent suburbs.

A final thought to leave you with is to consider the explosive growth in recent years of Wireless User Groups. These non-profit hobbyist metro Wifi networks are mostly used for fire-sharing and gaming (Internet access is specifically not provided). It shows that at least some South African consumers are willing to invest in their own telecoms infrastructure. Perhaps the way to bridge the gap between WUG and WISP is to form a fully licensed and professionally operated telecoms cooperative, like this one.

I replied with a detailed response, noting that the MX record points to a web virtual-hosting box at RSAWeb that isn’t accepting SMTP (I imagine having a broken MX suits Dynamic Seminars just fine), and that the website is a RSAWeb domain parking page.

Mark responded with this:

Thanks for putting this out in the wild, we do not condone spam & spammers. Unfortunately its always difficult to isolate these spammers as they tend to be small and difficult to detect unless we notice an abuse pattern or receive spam complaints! However we always make a concerted effort to stop these guys. Spam is inherintly evil and we have terminated the services of spammers that havent even made this ISPA hall of shame list (and will continue to do so).

He also said they’ve suspended the hosting until they have concluded their investigation. When I checked this afternoon, the domain parking page was still up, so I don’t know what they mean by “suspended”. Also, the specific spammer is number one out of twenty-five on the ISPA list, which hardly makes them small, or difficult to detect.

Nevertheless, RSAWeb at least took notice, which is probably more than I can say for the others in my ISPA members hall of shame.

1. The post started with the words “This morning”…
2. Couldn’t they get the list off the ISPA site and do their own lookups?

After filtering out domains without MX records, MX records without valid A records, and IP addresses for which whois lookups failed, the 63 listed domains whittled down to 44 servers, of which 14 are hosted by ISPA members.

So Localloop.co.za hereby presents the ISPA member hall of shame:

1. Hetzner (4 servers, 2 of which are hosted in Germany)¹
2. Webafrica (4 servers, 3 of which are hosted in the US)
3. Internet Solutions: (2 servers)²
4. Datapro (2 servers)
5. RSAWeb (1 server)³
6. AboutIT (1 server)

The only non-ISPA member is Intekom (Telkom SA) with one spammer mail server.

This does not prove the spammers listed use these mail servers for sending spam (they could use any other ISP for that, and judging by recent complaints on the IOZ list, Vodacom SA is one of the favorite providers for spammers).

Having said that, wouldn’t you expect ISPA members to uphold the spirit of the anti-spam clause in the ISPA code of conduct by not hosting mail servers for the ISPA’s own list of spammer domains?

1. Tie with Webafrica, but I put them first because one of WA’s spammer mail servers is hosted by Hertzner.
2. Tie with Datapro, but I put them first because AboutIT’s spammer mail server is hosted by IS.
3. Tie with AboutIT, but I put them first because they have their own data center.

SANReN will no doubt lead me to interesting discoveries that I hope to publish here¹ so please keep reading :-)

1. None of it will be confidential information. Also, none of the views expressed here represent any official position of my employers.

Having never worked in the ISP industry, I’m not really qualified to answer, but here goes anyway:

If you have this much bandwidth:

ZA can has traffic?

…then do this:

1. Randomly choose a price thats so expensive that it’s economically infeasible for your customers, then dial it back until you have ~100 reluctant customers.
2. Bill for transit by traffic volume in giga-bytes, call this a “quota”.
3. Hard-cap the “quota” so that the customer would exhaust it on the first day of month.
4. Expire (steal) the customer’s unused quota at the end of the month.
5. Oversubscribe like a lunatic, and then be vague about this to your customers.
6. Redefine “Internet access” to exclude certain popular protocols, and then buy a router that forwards those specific packets slower, or not at all.

If, on the other hand, you have this kind of bandwidth:

MOAR please.

…then do this:

1. Bill customers monthly for a 5-minute average traffic rate (Mbps).
2. Offer a variable per-Mbps discount if they commit to a minimum rate, and allow them to set the rate (to zero, if they like).
3. Don’t fine them for traffic spikes, instead, discard the top 5% or 10% of the 5-minute averages each month.

Here is the thing though, aren’t the graphs symptoms, and not causes of the pricing strategy?

So, back in the nineties, Telkom SA decided to call their ISP division “South African Internet Exchange”. That is also the last time SAIX updated their web site, but I digress. I think it’s possible that Telkom chose this name because from their PSTN perspective, an ISP looked like a telephone exchange for the Interwebs.

SAIX doesn’t pretend to be an IX (in fact, they don’t participate in any peering points in South Africa) so I’m willing to attribute their confusing name to misunderstanding.

However, the recently launched Africa Independent Network Exchange’s skilled professionals with many years of experience in the ISP industry should know that when you offer consumer broadband and hosting, you’re most definitely an ISP, so what’s with the name?

To be clear, their “peering service” is either IP transit (ISP business), or layer 2 circuits (telco business).

Failed Canine DDoS (note multiple attack vectors)

ph33r muh h4x0r sk1LLz

Motivated by frustration, our 5-year old dachshund launched a Distributed Denial of Service attack on our Internet connection, but failed (left). He made quick work of a nearby antenna feed coaxial cable, and he’s destroyed much tougher things, but he couldn’t nail this UTP cable…

I suspect the 24V DC on one of the copper pairs discouraged him. So could this be a solution for making network cabling less appetizing to the spectrum of gnawing and chewing creatures? Perhaps. But until it’s patented as anti-chew technology, who could be angry with a little furry being bearing a face like that :-)