Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


terrahost's new Nigeria location - benchmark & mini-review
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

terrahost's new Nigeria location - benchmark & mini-review

jsgjsg Member, Resident Benchmarker
edited February 2022 in Reviews

Originally this started as a "simple" review of the new @terrahost Nigeria location, but as it makes little sense to benchmark an almost empty node I don't make any statements on the node performance part (processor, memory, disk) other than confirming that, yes, you can expect the kind of performance you know from terrahost Norway; short, it's a really, really fast Ryzen 5k beast with a very fast NVMe.

What does make sense though and in fact is a question quite a few have in mind is how good (or bad) the new location's connectivity is. Here you go ... plus some detail, extra info (like route quality) and observations and thoughts:

Network Intro

While one might be surprised how actually not off the internet map Nigeria is with more than a handful of international cables (fiber) and yet another and fat one coming in 2023, after considering some relevant factors and doing some testing I can say with quite some confidence that terrahost NG is linked to London via 'Glo-1'. Interesting sidenote: Glo-1, unlike the vast majority of cables, seems not to be under us-american or european control but in the hands of a nigerian carrier and ISP, and is going directly to London (more precisely Bude, UK) with the exception of a short side-branch to Ghana (a neighbouring country).

So, I guess the puzzle of why everything from terrahosts nigerian location goes via London is solved. FWIW, it seems that terrahost made a really smart choice, both wrt the new location and the link.

Regarding the route info for each location I'm putting the first hops common to all targets right here:

[gateway].terrahost.com  < 1 ms
ipnxtelecoms, NG Lagos. 3 hops < 1 ms
ipnxtelecoms, London, 1 hop (direct via Glo-1)  93 ms
HE, London  95 ms

All the following route infos for each target are to be understood to follow this "common header". Also note a few extra possibly interesting targets.

Here are the results, based on a bit over 200 runs, halfway sorted by region:

 
US SJC softlayer [F: 0]
  DL [Mb/s]:      avg 47.4 - min 45.5 (96.1%), max 49.0 (103.4%)
  Ping [ms]:      avg 230.0 - min 226.2 (98.4%), max 233.2 (101.4%)
  Web ping [ms]:  avg 231.1 - min 226.3 (97.9%), max 237.7 (102.8%)

traceroute:
HE London, 2 hops   95 ms
[cross Atlantic]
[Entering networklayer.com network] 163 ms
networklayer.com Chicago, 2 hops  181 ms
networklayer.com Denver, 1 hop  202 ms
[Entering networklayer.com San Jose] 229 ms
networklayer.com San Jose, 6(!) hops  229 ms
- 19 hops total -
---------------------------------------------------------------------

US LAX clouvider [F: 48]
  DL [Mb/s]:      avg 35.2 - min 0.0 (0.0%), max 51.4 (145.9%)
  Ping [ms]:      avg 224.3 - min 0.0 (0.0%), max 232.5 (103.6%)
  Web ping [ms]:  avg 224.6 - min 0.0 (0.0%), max 240.1 (106.9%)

traceroute:
HE London, 2 hops   95 ms
[cross Atlantic]
HE New York, 1 hop 161 ms
HE Ashburn, 1 hop 164 ms
HE Los Angeles, 2 hops  220 ms
. . .
Clouvider LAX, 2 hops  229 ms
target  225 ms
- 16 hops total -
---------------------------------------------------------------------

US DAL leaseweb [F: 31]
  DL [Mb/s]:      avg 46.1 - min 0.0 (0.0%), max 58.1 (126.0%)
  Ping [ms]:      avg 194.6 - min 0.0 (0.0%), max 199.4 (102.5%)
  Web ping [ms]:  avg 195.6 - min 0.0 (0.0%), max 370.3 (189.4%)

traceroute:
HE, London,1 hop  96 ms
. . .
Zayo, London, 1 hop 96 ms
. . .
Zayo, Colorado, 1 hop 198 ms
. . . [3 hops]
Zayo, Dallas, 2 hops 198 ms
Leaseweb, Dallas, 2 hops  198 ms
target  198 ms
- 18 hops total -
---------------------------------------------------------------------

US NYC clouvider [F: 25]
  DL [Mb/s]:      avg 51.2 - min 0.0 (0.0%), max 70.5 (137.7%)
  Ping [ms]:      avg 161.7 - min 0.0 (0.0%), max 167.3 (103.4%)
  Web ping [ms]:  avg 169.0 - min 0.0 (0.0%), max 279.7 (165.5%)

traceroute:
HE, London, 1 hop  96 ms
Telia, London, 2 hops  96 ms
Telia, New York, 2 hops  163 - 180 ms
Clouvider, New York 7 hops  165 ms
target 165 ms
- 19 hops total -
---------------------------------------------------------------------

US WDC leaseweb [F: 15]
  DL [Mb/s]:      avg 58.0 - min 0.0 (0.0%), max 68.7 (118.6%)
  Ping [ms]:      avg 167.9 - min 167.4 (99.7%), max 188.8 (112.4%)
  Web ping [ms]:  avg 168.3 - min 167.5 (99.5%), max 188.8 (112.2%)

traceroute:
HE, London, 2 hops  96 ms
HE, New York, 1 hop 161 ms
HE, Ashburn, 3 hops, 166 ms
Leaseweb, Washington, 1 hop  166 ms
target  166 ms
- 14 hops total -
---------------------------------------------------------------------

BR PA ufpr.br [F: 0]
  DL [Mb/s]:      avg 32.2 - min 21.1 (65.4%), max 36.1 (111.9%)
  Ping [ms]:      avg 310.3 - min 0.0 (0.0%), max 335.1 (108.0%)
  Web ping [ms]:  avg 311.8 - min 0.0 (0.0%), max 353.6 (113.4%)

traceroute
HE, London, 2 hops  95 ms
HE, New York, 1 hop  161 ms
HE, Ashburn, 1 hop  164 ms
HE, Atlanta, 1 hop  179 ms
HE, Miami, 2 hops  197 ms
RNP (national edu network), Miami, 2 hops  196 ms
. . .
RNP, diverse, 5 hops ~ 315 ms
target  316 ms
- 22 hops total -
---------------------------------------------------------------------

BR SAO softlayer [F: 0]
  DL [Mb/s]:      avg 35.6 - min 19.7 (55.3%), max 41.1 (115.3%)
  Ping [ms]:      avg 271.5 - min 270.5 (99.6%), max 328.4 (121.0%)
  Web ping [ms]:  avg 275.8 - min 270.6 (98.1%), max 357.0 (129.4%)

traceroute
HE, London, 2 hops  95 ms
. . .
networklayer, London, 1 hop  96 ms
networklayer, New York, 1 hop  162 ms
. . .
sl-reverse.com, Sao Paolo, 3 hops  276 ms
target  27 ms
- 16 hops total -
[Note: both networklayer and sl-reverse seem to be softlayer (IBM) siblings]
---------------------------------------------------------------------

CL UNK ufro.cl [F: 1]
  DL [Mb/s]:      avg 33.7 - min 0.0 (0.0%), max 38.2 (113.3%)
  Ping [ms]:      avg 330.5 - min 327.7 (99.2%), max 355.5 (107.6%)
  Web ping [ms]:  avg 339.2 - min 327.7 (96.6%), max 638.2 (188.2%)

traceroute:
. . .
Telxius (Telefonica), London, 1 hop  95 ms
Telxius , ??,  170 ms
Telxius, Seattle,  195 ms
Telxius, Chile(?), 2 hops  304 ms
Gtd, Santiago, 2 hops,  307 ms
target  331 ms
- 15 hops total -
[Note: Telefonica seems to be quite secretive wrt. routing. Also, they seem to be tightly intertwined with Gtd]
---------------------------------------------------------------------

FR PAR testdebit [F: 0]
  DL [Mb/s]:      avg 98.4 - min 48.7 (49.5%), max 112.9 (114.7%)
  Ping [ms]:      avg 103.0 - min 101.8 (98.8%), max 106.0 (102.9%)
  Web ping [ms]:  avg 103.8 - min 102.0 (98.3%), max 106.0 (102.1%)

traceroute:
HE, London, 1 hop  95 ms
HE, Paris, 1 hop  106 ms
. . .
Bouygues, France, 3 hops  104 ms
target  104 ms
- 13 hops total -
---------------------------------------------------------------------


IT MIL softlayer [F: 0]
  DL [Mb/s]:      avg 95.2 - min 74.2 (77.9%), max 98.3 (103.2%)
  Ping [ms]:      avg 112.5 - min 0.0 (0.0%), max 127.8 (113.6%)
  Web ping [ms]:  avg 112.6 - min 0.0 (0.0%), max 127.8 (113.5%)

traceroute
HE, London, 2 hops  95 ms
. . .
networklayer, London, 1 hop  102 ms
. . . [2 hops]
networklayer, Milano, 2 hops  115 ms
target  114 ms
- 15 hops total -
---------------------------------------------------------------------


TR UNK 185.65.204.169 [F: 2]
  DL [Mb/s]:      avg 74.0 - min 0.0 (0.0%), max 77.3 (104.4%)
  Ping [ms]:      avg 142.3 - min 141.5 (99.5%), max 163.9 (115.2%)
  Web ping [ms]:  avg 142.6 - min 141.6 (99.3%), max 164.6 (115.4%)

traceroute
. . .
Telia, London, 1 hop  94 ms
. . .
Cogent, London, 1 hop  94 ms
Cogent, Amsterdam, 1 hop  101 ms
Cogent, Frankfurt, 1 hop  107 ms
Cogent, Munich, 1 hop  113 ms
Cogent, Vienna, 1 hop  119 ms
Cogent, Bratislava, 1 hop  120 ms
Cogent, Budapest, 1 hop  149 ms
Cogent, ???, 1 hop  127 ms
Cogent, Sofia, 1 hop  132 ms
Cogent, Istanbul, 2 hops  142 ms
binaryracks, Istanbul, 3 hops  143 ms
target  141 ms
- 24 hops total -
[Note (opinion): "Cogent - why use an express train when you can use a tramway stopping every 300 meters!"]
---------------------------------------------------------------------

GR UNK otenet.gr [F: 102]
  DL [Mb/s]:      avg 34.8 - min 0.0 (0.0%), max 79.5 (228.1%)
  Ping [ms]:      avg 81.0 - min 0.0 (0.0%), max 309.4 (382.0%)
  Web ping [ms]:  avg 85.4 - min 0.0 (0.0%), max 995.0 (1165.7%)

traceroute
HE, London, 1 hop  95 ms
HE, Prague, 1 hop  112 ms
HE, Vienna, 1 hop  118 ms
HE, Sofia, 1 hop  131 ms
HE, Athens, 1 hop  142 ms
. . .
OTEnet, Greece ?? hops  ca. 143 ms
17  * *^C
target  143 ms
- 16 to 20+ hops total -
[Note: This target is only here as a particularly nasty test, don't care, just remember to stay away from OTEnet.gr.
It's remarkable btw, that terrahost's Nigeria node actually has a success rate of about 50% with that target; that's an impressive result]
---------------------------------------------------------------------

NO OSL terrahost [F: 11]
  DL [Mb/s]:      avg 68.4 - min 0.0 (0.0%), max 92.1 (134.8%)
  Ping [ms]:      avg 122.0 - min 0.0 (0.0%), max 126.6 (103.7%)
  Web ping [ms]:  avg 122.3 - min 0.0 (0.0%), max 144.5 (118.1%)

traceroute:
HE, London, 1 hop  95 ms
HE, Amsterdam, 1 hop  113 ms
HE, Stockholm, 1 hop  120 ms
. . .
terrahost, 2 hops  123 ms
target  125 ms
- 13 hops total -
---------------------------------------------------------------------

IS UNK system.is [F: 0]
  DL [Mb/s]:      avg 80.0 - min 29.8 (37.3%), max 86.5 (108.1%)
  Ping [ms]:      avg 133.1 - min 0.0 (0.0%), max 143.4 (107.7%)
  Web ping [ms]:  avg 133.4 - min 0.0 (0.0%), max 144.3 (108.2%)

traceroute:
HE, London, 1 hop  95 ms
HE, Amsterdam, 1 hop  101 ms
. . .
simnet.is, Iceland, 1 hop  143 ms
Gagnaveita, Iceland, 1 hop  141 ms
target  143 ms
- 12 hops total -
---------------------------------------------------------------------

RU MOS hostkey [F: 0]
  DL [Mb/s]:      avg 64.9 - min 21.0 (32.4%), max 82.8 (127.6%)
  Ping [ms]:      avg 139.4 - min 0.0 (0.0%), max 142.7 (102.4%)
  Web ping [ms]:  avg 139.4 - min 0.0 (0.0%), max 142.9 (102.5%)

traceroute:
HE, London, 1 hop  95 ms
HE, Amsterdam, 1 hop  102 ms
HE, tockholm, 1 hop  120 ms
. . . [2 hops]
target  142 ms
- 12 hops total -
---------------------------------------------------------------------

RO BUC 185.183.99.8 [F: 0]
  DL [Mb/s]:      avg 77.8 - min 22.8 (29.3%), max 88.1 (113.3%)
  Ping [ms]:      avg 129.0 - min 0.0 (0.0%), max 171.4 (132.9%)
  Web ping [ms]:  avg 129.6 - min 0.0 (0.0%), max 171.4 (132.3%)

traceroute:
HE, London, 1 hop  95 ms
HE, Manchester, 1 hop  99 ms
. . .
M247, London, 1 hop,  101 ms
M247, "ddos-protected-ip"  101 ms
. . . [7 hops]
target  135 ms
- 19 hops total -
---------------------------------------------------------------------

NL AMS leaseweb [F: 0]
  DL [Mb/s]:      avg 102.4 - min 25.9 (25.3%), max 113.6 (110.9%)
  Ping [ms]:      avg 97.9 - min 0.0 (0.0%), max 102.0 (104.2%)
  Web ping [ms]:  avg 98.4 - min 0.0 (0.0%), max 104.1 (105.8%)

traceroute:
HE, London, 2 hops  95 ms
. . . [ 4 hops]
target  100 ms
- 13 hops total -
---------------------------------------------------------------------

DE FRA core-backbone [F: 29]
  DL [Mb/s]:      avg 77.8 - min 0.0 (0.0%), max 106.8 (137.2%)
  Ping [ms]:      avg 104.7 - min 0.0 (0.0%), max 106.9 (102.1%)
  Web ping [ms]:  avg 105.2 - min 0.0 (0.0%), max 108.3 (102.9%)

traceroute
HE, London, 2 hops  95 ms
. . .
core-backbone, Amsterdam, 1 hop  99 ms
core-backbone, Frankfurt, 1 hop  106 ms
target  106 ms
- 12 hops total -
---------------------------------------------------------------------

UK LON clouvider [F: 1]
  DL [Mb/s]:      avg 90.9 - min 0.0 (0.0%), max 122.0 (134.2%)
  Ping [ms]:      avg 92.0 - min 0.0 (0.0%), max 95.8 (104.1%)
  Web ping [ms]:  avg 92.1 - min 0.0 (0.0%), max 96.0 (104.3%)

traceroute:
HE, London, 2 hops  95 ms
. . . [7 hops]
Clouvider, London, 2 hops  94 ms
target  95 ms
- 18 hops total -
---------------------------------------------------------------------

AU MEL dediserve [F: 6]
  DL [Mb/s]:      avg 25.8 - min 0.0 (0.0%), max 28.6 (110.9%)
  Ping [ms]:      avg 389.6 - min 0.0 (0.0%), max 401.5 (103.1%)
  Web ping [ms]:  avg 390.8 - min 0.0 (0.0%), max 405.9 (103.9%)

traceroute:
HE, London, 2 hops  95 ms
HE, New York, 1 hop  161 ms
HE, Ashburn, 1 hops  175 ms
HE, Los Angeles, 1 hops  222 ms
. . .
coloau, "Los Angeles" (actually AU), 1 hop  391 ms
coloau, NW, 1 hop  390 ms
coloau, Melbourne, 2 hops  389 ms
. . . [between 1 and countless hops]
target  391 ms
- 17 to 20+ hops total -
---------------------------------------------------------------------

JP TOK linode [F: 35]
  DL [Mb/s]:      avg 29.4 - min 0.0 (0.0%), max 38.5 (130.7%)
  Ping [ms]:      avg 298.9 - min 297.8 (99.6%), max 302.5 (101.2%)
  Web ping [ms]:  avg 299.4 - min 297.8 (99.5%), max 304.4 (101.7%)

traceroute:
HE, London, 1 hop  95 ms
HE, New York, 1 hop  159 ms
HE, Chicago, 2 hops  180 ms
HE, Seattle, 1 hops  218 ms
HE, Tokyo, 3 hops  301 ms
Linode, Tokyo, 1 hop  301 ms
target  302 ms
- 16 hops total -
---------------------------------------------------------------------

CN UNK cqu.edu.cn [F: 1]
  DL [Mb/s]:      avg 27.5 - min 0.0 (0.0%), max 29.5 (107.5%)
  Ping [ms]:      avg 382.4 - min 379.8 (99.3%), max 395.3 (103.4%)
  Web ping [ms]:  avg 386.0 - min 379.8 (98.4%), max 537.1 (139.1%)

traceroute:
HE, London, 1 hop  95 ms
HE, Paris, 1 hop  104 ms
HE, Marseille, 1 hop  111 ms
HE, Singapore, 1 hop  250 ms
HE, Hongkong, 2 hops  282 ms
. . .
CN Edu&Research network, 1 hop  350 ms
. . .
CN Edu&Research network, 7 hops  384 ms [some hopping in circles]
. . . [11 hops]
^C [I won't wait all day for the chinese intelligentsia (a joke with a smirk) to end
their funny routing loops]
- 24 to 30+ hops total [Bring lots of free time along ...] -
---------------------------------------------------------------------

CN HK leaseweb [F: 29]
  DL [Mb/s]:      avg 33.7 - min 0.0 (0.0%), max 42.0 (124.4%)
  Ping [ms]:      avg 274.4 - min 0.0 (0.0%), max 292.3 (106.5%)
  Web ping [ms]:  avg 280.5 - min 0.0 (0.0%), max 294.6 (105.0%)

traceroute:
HE, London, 1 hop  100 ms
HE, Paris, 1 hop  104 ms
HE, Marseille, 1 hop  111 ms
HE, Singapore, 1 hop  250 ms
HE, Hongkong, 2 hops  280 ms
. . . [2 hops]
leaseweb, Hongkong, 2 hops  287 ms
target  276 ms
- 17 hops total -
---------------------------------------------------------------------

CN HK mirror.rackspace.hk [F: 6]
  [results hidden]

[This is actually *not* a server in Hongkong but one on the us-american East Coast.
Which means that Rackspace are playing - very dirty - games with dozens of distros 
and major foss projects!]
- no hops for liars and scammers -
---------------------------------------------------------------------

SG SGP leaseweb [F: 38]
  DL [Mb/s]:      avg 36.1 - min 0.0 (0.0%), max 46.5 (129.0%)
  Ping [ms]:      avg 250.0 - min 0.0 (0.0%), max 286.6 (114.6%)
  Web ping [ms]:  avg 250.7 - min 0.0 (0.0%), max 295.5 (117.9%)

traceroute:
HE, London, 1 hop  95 ms
HE, Paris, 1 hop  107 ms
HE, Marseille, 1 hop  112 ms
HE, Singapore, 1 hop  250 ms
. . . [2 hops]
leaseweb, Singapore, 2 hops  245 ms
target  255 ms [Uhm, losing 10 ms within SGP?]
- 15 hops total -
---------------------------------------------------------------------

IN MUM webwerks [F: 0]
  DL [Mb/s]:      avg 42.6 - min 24.6 (57.8%), max 50.9 (119.4%)
  Ping [ms]:      avg 249.6 - min 0.0 (0.0%), max 268.7 (107.6%)
  Web ping [ms]:  avg 255.7 - min 0.0 (0.0%), max 322.7 (126.2%)

traceroute:
HE, London, 3 hops  94 ms
Vodafone, London, 3 hops  119 ms
. . . [2 hops]
Vodafone, India  223 ms
target  225 ms
- 16 hops total -

My resumé:

As I already hinted earlier I'm expecting this new terrahost location to attract lots of customers, both from outside and inside Nigeria. Simple reason: while routes with 95 ms added in my seem a bit mee to us European, North-Americans, etc. terrahosts products there highly likely is THE entry and exit port to/fro Africa, and looking at the growing fiber capacity and the economic efforts of the whole region one doesn't need to be a genius to recognize that terrahost is sitting on a quite exclusive (so far) and promising "mine".

While I would have preferred to see way more twelfe99 in my tests of the terrahost VPS I also see that, especially considering the location, not going for the the top spot but for decent quality at an excellent price was the right decision. And I can't complain, HE gave me little reason to be unhappy, their part of the routes is rarely a problem - unlike for example Cogent who quite reliably f_cks routing up, surpassed only by the chinese government's reasearch and edu network which leads me to believe that some important chinese official thinks that higher numbers, both in time and in hop count are better ...
(No, I'm not at all anti-China, but when a country reliably f_cks up in some really important aspect I call a spade a spade)

Back to terrahost: kudos, @RobMonster and terrahost team, your thinking was smart, your plan was smart, and your implementation is well done. And your routing isn't bad at all either and I think your decision to go with HE where it makes sense was a good one (and a friendly smile to HE from me, you've got quite decent ;) )

And now, why are you wasting time reading my rumblings instead of looking at terrahost's shop? What? "Not yet available" you say? Rob Monster and terrahost team, MOVE! Open the doors to the shop already! Or do you want me to send you a picture of my face to motivate you? (Trust me, there are less horrifying nightmares than seeing my face ...) *g

Comments

  • edited February 2022

    Pretty nice review, thanks! The Glo-1 line indeed explains a lot, and is an interesting and pretty unique choice. Twisted my arm into getting one in the near future lol.

    Thanked by 1jsg
  • peenus

    Thanked by 1risharde
  • Does "resumé" mean something different in Europe than North America? I was expecting a listing of prior experience, education and skills, etc. I'm guessing by the rant after, he more meant "in conclusion"?

  • @TimboJones said:
    Does "resumé" mean something different in Europe than North America? I was expecting a listing of prior experience, education and skills, etc. I'm guessing by the rant after, he more meant "in conclusion"?

    "résumé" means "summary" in French. English speaking countries use it to say "my CV" because it's a summary of one's work experience and history.

    Thanked by 2jsg TimboJones
  • Use cases may differ but for me I would have loved to see a network review regarding (Western) Africa locations. I wouldn't go for a Lagos location and than see how good the connection to Japan/US or elsewhere far away is but want to know the quality for neighboring countries etc.

    Some unqualified tests of mine show a lot of connections going London (or Paris) first and than back which would be worse than using a London location for (Western) Africa.

    So a test with ISPs in the region would be nice if anyone can come up with that.

  • @OhJohn said:
    So a test with ISPs in the region would be nice if anyone can come up with that.

    I created a traceroute measurement from Africa and the middle east.

    IPv6: https://atlas.ripe.net/measurements/38978404/#probes
    IPv4: https://atlas.ripe.net/measurements/38978419/#probes

    Thanked by 1OhJohn
  • OhJohnOhJohn Member
    edited February 2022

    That's what I meant, Cameroon, Ghana, Togo etc. on ipv4 all using the London route. At least South Sudan seems to have a rather direct route :)

  • @OhJohn said:
    That's what I meant, Cameroon, Ghana, Togo etc. on ipv4 all using the London route. At least South Sudan seems to have a rather direct route :)

    And this is common in basically all regions except US and EU. Local ISPs have peering agreements that may not be so good for two reasons, to save costs and to try to sell their transit. Even if a ISP shows good ping to a specific datacenter, maybe the person running the RIPE atlas probe just struck the routing lottery and it's no guarantee.

    tl;dr Unless you are paying a lot, this is not really the server provider's problem.

  • Rob mounted terrahost software on his hardware. In this case locations should increase and be equal to epiks.

  • jsgjsg Member, Resident Benchmarker

    @OhJohn said:
    ... I would have loved to see a network review regarding (Western) Africa locations.

    So a test with ISPs in the region would be nice if anyone can come up with that.

    The problem is test targets. Sadly there are whole countries for which I don't have a single reliably reachable and working test target. Plus the same is true the other way around; some targets working fine from location/country A don't work from location/country B.
    Another limiting factor might be the lack of fat fibers and/or the (high) price for traffic in some regions which means there aren't many ISPs/providers who are providing 100+ MB test files and the traffic needed for downloading.

    If you can tell me (PM or here) some african test targets (http, please, not httpS) that work reliably, I'll ask terrahost whether I can have access to the test VPS for a couple more days and test those targets.

    Thanked by 1OhJohn
  • jsgjsg Member, Resident Benchmarker

    @plastik said:

    @TimboJones said:
    Does "resumé" mean something different in Europe than North America? I was expecting a listing of prior experience, education and skills, etc. I'm guessing by the rant after, he more meant "in conclusion"?

    "résumé" means "summary" in French. English speaking countries use it to say "my CV" because it's a summary of one's work experience and history.

    Yes, and when we mean a CV we simply say CV (curriculum vitae).

  • RobMonsterRobMonster Member, Host Rep
    edited February 2022

    Thanks @jsg - this was a great writeup.

    For those not aware, Nigeria is the most populous nation in the African continent. The network routing across Africa also made Lagos the preferred location for serving the continent with local infrastructure.

    Looking ahead, I expect that our BitMitigate.com division will soon have CDN nodes online for better serving Africa, starting in Lagos and expanding from there to decentralized CDN nodes for cost-effective decentralized cacheing of content.

    Looking a bit further ahead, we have been incubating a hyper-decentralized network infrastructure solution called Toki.com. We are looking to further accelerate this development in 2022 for a range of decentralized and affordable configurations.

  • NoCommentNoComment Member
    edited February 2022

    @jsg said:

    @OhJohn said:
    ... I would have loved to see a network review regarding (Western) Africa locations.

    So a test with ISPs in the region would be nice if anyone can come up with that.

    The problem is test targets. Sadly there are whole countries for which I don't have a single reliably reachable and working test target. Plus the same is true the other way around; some targets working fine from location/country A don't work from location/country B.
    Another limiting factor might be the lack of fat fibers and/or the (high) price for traffic in some regions which means there aren't many ISPs/providers who are providing 100+ MB test files and the traffic needed for downloading.

    If you can tell me (PM or here) some african test targets (http, please, not httpS) that work reliably, I'll ask terrahost whether I can have access to the test VPS for a couple more days and test those targets.

    I think @OhJohn is referring more specifically to consumer ISPs, so the routing you would get serving end-users. It does not matter if many telcos/consumer ISPs have specific routes/peers that have very good routing. These are the minority and most users will have to deal with the poor routing to various parts of the world. From africa, it's obvious these ISPs will have a lot of transit to somewhere in west EU. It's just cheap that way. For east asia, it would be west US. This is just how consumer ISPs operate in these regions.

  • DPDP Administrator, The Domain Guy

    I hope Asia’s next - preferably 🇸🇬 😁

    Thanked by 1dosai
  • NeoonNeoon Community Contributor, Veteran

    @babuum said:

    @OhJohn said:
    So a test with ISPs in the region would be nice if anyone can come up with that.

    I created a traceroute measurement from Africa and the middle east.

    IPv6: https://atlas.ripe.net/measurements/38978404/#probes
    IPv4: https://atlas.ripe.net/measurements/38978419/#probes

    Yea, the routing to/from Europe is good, plus local routing inside the country, outside the country it seems to be bad.

  • jsgjsg Member, Resident Benchmarker

    Quick update: terrahost graciously prolonged my access.

    So, if you want me to test other targets, just let me know, be it here or via PM.
    @OhJohn et al ...

  • @plastik said:

    @TimboJones said:
    Does "resumé" mean something different in Europe than North America? I was expecting a listing of prior experience, education and skills, etc. I'm guessing by the rant after, he more meant "in conclusion"?

    "résumé" means "summary" in French. English speaking countries use it to say "my CV" because it's a summary of one's work experience and history.

    Oof, I should have known that. Thanks

Sign In or Register to comment.