Type Nairobi in the search box, and you'll see that Akamai has a presence in Nairobi. By the way also in Kampala (Uganda) and Kigali (Rwanda): http://wwwnui.akamai.com/gnet/globe/index.html
Why the Internet in Sub-Saharan Africa Sucks
71–80 of 127 posts
Re: Why the Internet in Sub-Saharan Africa Sucks
#72We are currently building applications for our shops in Africa (Rwanda, Kenya and Uganda) at my current company. Kenya is not even the worst of all three, Uganda is really a disaster in terms of connectivity. In Uganda it's even difficult to get actual phone signal in most of the country. And you never have reliable Internet anyway, it works like half of the time even if you are in the few right places in the country…
That sounds like a challenge! What line of business are you in that can survive that bad connectivity?
We are selling small solar panel kits in small towns on those three countries. The goal is to have a few lights in places where there is no electricity grid at all (it's a for-profit company, not charity). We handle pretty much everything ourselves there since there is not a lot of services pre-built to help you. The shops are selling the kits on payplan since it's quite expensive for the local purchase power. We are the equivalent of a car dealer locally to give a comparison. And the shops are running some software to handle all of that, so this is where my problems are starting ! :)
Re: Why the Internet in Sub-Saharan Africa Sucks
#73Taking one second to establish the connection doesn't really explain a 12 second load time. TCP takes a little bit to ramp up, but not 30 round trips. Plus you have TCP preconnect as a lightweight way to avoid that initial delay.
However, looking at Twitters front page I see it establishing connections to 6 different domains, that's one TLS session for each one.
Re: Why the Internet in Sub-Saharan Africa Sucks
#74In one of Google's lobby there is a 3-D animation of a spinning globe, with spikes that appear wherever and whenever searches are performed, with the height of the spikes being proportional to the rate of query submissions. The recruiter who showed me this was very proud of google's accomplishment and indeed it was impressive to see the spikes erupting from the east and west coast of the US, Europe, Japan and India.…
> There were modest spikes in south africa and brazil. Not many anywhere else.
Huh? South America has good overall Internet penetration[1] and Chile is going to 70% right now.
[1] https://en.wikipedia.org/wiki/List_of_countries_by_number_of...
Re: Why the Internet in Sub-Saharan Africa Sucks
#75Earlier quoted context omitted.
JHB to Europe is generally round the 190ms mark. 170 if you're closer to the undersea cables. No way you're going to drop that to 100ms.
JHB is much further away from Europe than Kenya, though. JHB means usually going to Cape Town first, then up the west coast of the continent, wheras Mombassa uses SeaCom, going around the east of the continent. That saves a lot of time from a pure distance perspective alone.
>JHB is much further away from Europe than Kenya, though.
Indeed. I was responding to a comment that was discussing JHB though.
SA traffic is split - some of it goes round east (Seacom), some west (SAT3). Ends up being roughly 200ms east or west.
Source: Got my interwebs over said cables for years
Re: Why the Internet in Sub-Saharan Africa Sucks
#76In one of Google's lobby there is a 3-D animation of a spinning globe, with spikes that appear wherever and whenever searches are performed, with the height of the spikes being proportional to the rate of query submissions. The recruiter who showed me this was very proud of google's accomplishment and indeed it was impressive to see the spikes erupting from the east and west coast of the US, Europe, Japan and India.…
Google isn't the leading search engine in all countries of the world, so they should not spike everywhere. I'm surprised you didn't notice Russia and China not spiking as much either.
Google is the leading search engine in both Africa and South America though.
Re: Why the Internet in Sub-Saharan Africa Sucks
#77We have contemplated a POP in Kenya. We will get there. Johannesburg should be less than 200ms, so let me see if I can improve that... I would expect about 100ms.
JHB to Europe is generally round the 190ms mark. 170 if you're closer to the undersea cables. No way you're going to drop that to 100ms.
Who is talking about Europe? This is about reaching Kenya. It should actually be ~60-70ms.
Re: Why the Internet in Sub-Saharan Africa Sucks
#78We have contemplated a POP in Kenya. We will get there. Johannesburg should be less than 200ms, so let me see if I can improve that... I would expect about 100ms.
JHB to Europe is generally round the 190ms mark. 170 if you're closer to the undersea cables. No way you're going to drop that to 100ms.
Who is talking about Europe? This is about reaching Kenya. It should actually be ~60-70ms.
Re: Why the Internet in Sub-Saharan Africa Sucks
#79Earlier quoted context omitted.
It's only in Chrome and only supported by Google's web server. They are using it though. You can see the stats by going to chrome://net-internals/#quic They are "working" to get it into Apache Traffic Server.
Yeah, its communicating over port 443. Apache Traffic Server would sure be a big step forward, if and when that happens.
443 UDP though - different to anything else.
Re: Why the Internet in Sub-Saharan Africa Sucks
#80Earlier quoted context omitted.
I'm in Africa and... I'm sorry, but we have nameservers already. That's not really a problem. Are you really talking about DNS servers here? I wasn't sure. The issues which prevent those spikes you'd like to see on the maps at Google are education (primarily literacy, especially English literacy), access to devices, and culture. Access to devices is being solved by cheap smartphones becoming obtainable to some degree…
> consider someone in a village (with a cellphone and access to a data network) who wants to find out why something happened to the carrot he was growing at home. All of us would pop open Google and search about some information about carrots. A guy in a village would head up the road to talk to his friend who's been farming for 50 years and knows about carrots Speaking as a US native, the reason we'd look for carrot…
Yes, but this is the central advantage of distributed information: while your carrot-farmer buddy is a higher-bandwidth and possibly deeper source of carrot information than google is, that guy doesn't know so much about goat husbandry. Or fishing. Or roof construction, or car repair, etc.
So if carrot info is all you ever need, great -- but otherwise, you are going to need nearby buddies who are experts in many, many other things. In fact, you're going to need far more buddies than you can ever actually have, due to the limits of the number of social connections any of us can have (not to mention the limits that time itself imposes on the conversations we'd have to have in order to maintain those relationships even if we could keep them all straight in our heads).
Google's primary utility isn't in making you a perfect expert on anything in particular -- it's that it allows you to become sufficiently expert at anything at all in a really short period of time. Even an army of friends, with the most varied experience you can imagine, cannot come close to matching the breadth of expertise available via the net.