Afrihost - Pure Fibre Feedback Thread Part 2

Satedah said:
Net down again.
W
T
F

Same issue as a few weeks ago that was supposedly fixed.

What is going on. Trying to watch the footy here.

Is constant disconnects every day just a thing now
Click to expand...
I'm starting to think its our area. Always when I come to check when my net has died, you've already posted.
 
sand_man said:
Apple podcasts buffering, on a gigabit account. Frigging podcasts!! This boils my blood.
Click to expand...
Yeah. Issue is whoever hosts rugbybits pod cast. Whatever servers. I don't know how this s hit works but other pods are playing fine. The grade cricketer with a server probably in Australia is perfect. Rugbybits a local podcaster probably hosted locally and total s hit. What gives?
 
After the two cable breaks, I'm not sure which route traffic is taking (I'm guessing Equiano) but < 140ms is a record for me, the lowest I've ever got before this was 170ms



20230812081609_1.jpg
 
alph4byt3 said:
After the two cable breaks, I'm not sure which route traffic is taking (I'm guessing Equiano) but < 140ms is a record for me, the lowest I've ever got before this was 170ms



View attachment 1570936
Click to expand...
You're connected to Bahrain CISP been using Equiano for a couple days and my ping to EU is the same as WACS from CT to UK is 142

On WiFi on speedtest
 

Attachments

  • Screenshot_20230812_105524_Speedtest.jpg
    Screenshot_20230812_105524_Speedtest.jpg
    86.7 KB · Views: 31
AfriNatic said:
It comes at a cost though. Higher latency to EU West like London and the US and even higher to South America.
Click to expand...
I'm really curious how expensive SACS capacity is given the latency reductions it offers since it's part of the comparison here. Testing from a box connected to TENET (which has SACS capacity) I get 161ms latency to an server in Florida (which fits in line with the figure here). Same box on my Afrihost connection gives 321ms latency (i.e. twice as high).

SACS plus current EU configuration would seem to give:
* Lower latency to a decent chunk of the EU, at a cost of slight increase in latency to EU West (not sure if for the EU as a whole this is on net positive/breakeven/negative)
* Cut latency to South America maybe by 3/4
* Cut latency to southern US by maybe 33-50%
* Rest of North America probably pretty comparable to normal conditions.

Guessing that transit on SACS must be substantially more expensive (or locked into long-term contracts related to North Atlantic capacity).
 
crashingintoyou said:
I'm really curious how expensive SACS capacity is given the latency reductions it offers since it's part of the comparison here. Testing from a box connected to TENET (which has SACS capacity) I get 161ms latency to an server in Florida (which fits in line with the figure here). Same box on my Afrihost connection gives 321ms latency (i.e. twice as high).

SACS plus current EU configuration would seem to give:
* Lower latency to a decent chunk of the EU, at a cost of slight increase in latency to EU West (not sure if for the EU as a whole this is on net positive/breakeven/negative)
* Cut latency to South America maybe by 3/4
* Cut latency to southern US by maybe 33-50%
* Rest of North America probably pretty comparable to normal conditions.

Guessing that transit on SACS must be substantially more expensive (or locked into long-term contracts related to North Atlantic capacity).
Click to expand...
The bigger problem that comes with having SACS capacity is certain updates and services suddenly start pulling from those servers in those regions and it maxes out the capacity.

Coo Ideas had SACS capacity when I first joined them and whenever any Blizzard game was released it would saturate their capacity and they would need to turn it off.

Throw in most game servers these days are Google/AWS/Microsoft they all pickup locally so the actual gain is minimal as its only direct hosts you could use the transit with. To fully utilise this capacity and go to all the servers you would need a VPN solution or to de peer all the big guys locally which would come at a huge cost since you would need to peer with them internationally.

You also then get the issue where certain games you can't change matchmaking server and it defaults your gamers onto the lowest latency server which would be the likes of a South America server which has extremely low population thus high queue times or no games.

I have a friend who was with Webafrica and their SMITE traffic went CT->JHB->Singapore->EU so a similar thing happens when having SACS capacity
 
crashingintoyou said:
I'm really curious how expensive SACS capacity is given the latency reductions it offers since it's part of the comparison here. Testing from a box connected to TENET (which has SACS capacity) I get 161ms latency to an server in Florida (which fits in line with the figure here). Same box on my Afrihost connection gives 321ms latency (i.e. twice as high).

SACS plus current EU configuration would seem to give:
* Lower latency to a decent chunk of the EU, at a cost of slight increase in latency to EU West (not sure if for the EU as a whole this is on net positive/breakeven/negative)
* Cut latency to South America maybe by 3/4
* Cut latency to southern US by maybe 33-50%
* Rest of North America probably pretty comparable to normal conditions.

Guessing that transit on SACS must be substantially more expensive (or locked into long-term contracts related to North Atlantic capacity).
Click to expand...

I always hope to see SACS be used in the future for reduced latency both to North and South America. In fact the reason I created an account here was to ask about it.
 
CrypticZA said:
You're connected to Bahrain CISP been using Equiano for a couple days and my ping to EU is the same as WACS from CT to UK is 142

On WiFi on speedtest
Click to expand...

well there goes my excitement and happiness on low EU latency, no wonder why the junkrat was speaking arabic
 
alph4byt3 said:
well there goes my excitement and happiness on low EU latency, no wonder why the junkrat was speaking arabic
Click to expand...
I think its control shift N shows net graph in game top left shows the server region will be like AMS for Amsterdam
 
CrypticZA said:
The bigger problem that comes with having SACS capacity is certain updates and services suddenly start pulling from those servers in those regions and it maxes out the capacity.

Coo Ideas had SACS capacity when I first joined them and whenever any Blizzard game was released it would saturate their capacity and they would need to turn it off.

Throw in most game servers these days are Google/AWS/Microsoft they all pickup locally so the actual gain is minimal as its only direct hosts you could use the transit with. To fully utilise this capacity and go to all the servers you would need a VPN solution or to de peer all the big guys locally which would come at a huge cost since you would need to peer with them internationally.

You also then get the issue where certain games you can't change matchmaking server and it defaults your gamers onto the lowest latency server which would be the likes of a South America server which has extremely low population thus high queue times or no games.

I have a friend who was with Webafrica and their SMITE traffic went CT->JHB->Singapore->EU so a similar thing happens when having SACS capacity
Click to expand...

Afrihost created PPTP VPN servers to help with a more direct route to EU as we know from their thread about it. Would it be a possibility if they invested the funds into it to create one that makes use of the SACS route? @AfriNatic would actually like to hear your thoughts on this.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X