Azure in SA: Here’s why the new South African Datacenters are so good.
How much of a difference does it make to have a local Microsoft Azure Datacentre?
When Microsoft announced that they were going to develop local datacentres in South Africa, not only was it going to be awesome for local clients, it would almost give us a much-needed boost to the economy. Local service providers, such as ourselves, would now be given the world-class speeds and experience our peers over the ponds have been experiencing for years.
Now that the datacentres have finally arrived, I wanted to do a few tests to compare the experience and the speed of a local datacentre versus a remote datacentre. I used the United Kingdom South datacentre as an option because this was our go-to location in the past, as it yielded the best speed and experience results for us here in South Africa.
I decided to kick things off and start my tests. I built a quick Virtual Machine in SA North, which took just over 6 minutes to be fully deployed and made sure it was ready with Windows Server 2019.
Once that was successfully deployed, I mirrored the same build using UK South Resource. The completed build was slightly little quicker and it was completed at 3 minutes and 48 seconds.
The first win goes to the UK but I’m not ruling out South Africa just yet. This could be because of it being launch day and hundreds of people are trying this out right now.
Now, we will look at the machines side-by-side starting with a Ping.
I’m going to run a ping from my office in Hillcrest, Kwa-Zulu Natal, using our 100mb fibre business line.
First the SA VM:
64 bytes from 18.104.22.168: icmp_seq=63 ttl=112 time=32.6 ms 64 bytes from 22.214.171.124: icmp_seq=64 ttl=112 time=32.7 ms 64 bytes from 126.96.36.199: icmp_seq=65 ttl=112 time=32.6 ms 64 bytes from 188.8.131.52: icmp_seq=66 ttl=112 time=32.8 ms 64 bytes from 184.108.40.206: icmp_seq=67 ttl=112 time=33.1 ms 64 bytes from 220.127.116.11: icmp_seq=68 ttl=112 time=32.9 ms 64 bytes from 18.104.22.168: icmp_seq=69 ttl=112 time=33.4 ms 64 bytes from 22.214.171.124: icmp_seq=70 ttl=112 time=32.8 ms 64 bytes from 126.96.36.199: icmp_seq=71 ttl=112 time=33.1 ms 64 bytes from 188.8.131.52: icmp_seq=72 ttl=112 time=32.6 ms 64 bytes from 184.108.40.206: icmp_seq=73 ttl=112 time=32.4 ms 64 bytes from 220.127.116.11: icmp_seq=74 ttl=112 time=32.5 ms 64 bytes from 18.104.22.168: icmp_seq=75 ttl=112 time=33.2 ms 64 bytes from 22.214.171.124: icmp_seq=76 ttl=112 time=32.9 ms
Now the UK VM;
64 bytes from 126.96.36.199: icmp_seq=74 ttl=114 time=175 ms 64 bytes from 188.8.131.52: icmp_seq=75 ttl=114 time=176 ms 64 bytes from 184.108.40.206: icmp_seq=76 ttl=114 time=175 ms 64 bytes from 220.127.116.11: icmp_seq=77 ttl=114 time=176 ms 64 bytes from 18.104.22.168: icmp_seq=78 ttl=114 time=174 ms 64 bytes from 22.214.171.124: icmp_seq=79 ttl=114 time=174 ms 64 bytes from 126.96.36.199: icmp_seq=80 ttl=114 time=175 ms 64 bytes from 188.8.131.52: icmp_seq=81 ttl=114 time=175 ms 64 bytes from 184.108.40.206: icmp_seq=82 ttl=114 time=176 ms 64 bytes from 220.127.116.11: icmp_seq=83 ttl=114 time=175 ms 64 bytes from 18.104.22.168: icmp_seq=84 ttl=114 time=175 ms 64 bytes from 22.214.171.124: icmp_seq=85 ttl=114 time=175 ms 64 bytes from 126.96.36.199: icmp_seq=86 ttl=114 time=175 ms 64 bytes from 188.8.131.52: icmp_seq=87 ttl=114 time=175 ms
As expected, the SA machine obliterates the UK machine by 130ms on average.
It goes without saying, this is a clear, obvious and important win for the SA Data Centre. This can obviously be improved with the position and tracerouting of your ISP breakout.
Now for a less obvious test, who has faster, greater bandwidth.
First up is South Africa. The results speak for themselves. Massive Ping, Download and monumental upload speed. That upload speed is something we have dreamed about for years and we are finally seeing it come to light.
The UK struggles to come anywhere close to those upload speeds but it’s still no slouch. I wouldn’t throw those results out of bed if you know what I mean.
Now, this really surprised me. South African datacentres won on bandwidth. That is a welcome surprise. As I’ve said before, this is something I’ve dreamed about for years and to see it come to life is truly spectacular.
While having great ping and amazing bandwidth is important, if your experience is bad, those results wouldn’t matter.
I’m going to perform the exact same clicks and page loads on an RDP session to see what the results are. You can view my test below.
As you can guess from the bandwidth and ping results, the SA RDP experience is clean and tidy with very little lag whereas the UK experience while still good, is slower and much greater lag overall.
As we’re nearing the end of the test, we need to cover another important topic. Cost. Which is more? Well this is another welcome surprise. The UK cost, which is around the average cost globally, is $214,62 (UK Cost)
The SA Cost is $178,85, which is around R500 a month saving, for the exact same, faster VM.
The only cheaper options were North Korea at $168,63 , North Europe $169,36 South Central, West Central and East US 2 $170,82 and from South African, if we’re being truly honest, the lag to those locations is far worse than the UK South, which is the closest node off our international traffic backbone.
So what’s the verdict so far?
Well that’s easy. As we say in sunny, beautiful South Africa say, local is lekker. But it doesn’t stop there, local is lekker, faster and cheaper. The South African Microsoft Azure datacentres have really impressed me with their speed, cost and overall performance. I’m really looking forward to utilizing this amazing service and enabling productivity with you in your business. Check back soon for more info here on our blog.