Spark UFB connection same city latency

Forums › Spark New Zealand (including Skinny and BigPipe) › Spark UFB connection same city latency

nbroad

320 posts

Ultimate Geek
+1 received by user: 39

#285909 24-May-2021 15:39

I have two Spark Business UFB MAX connections in Palmerston North.

Both connections get 3 ms RTT and 900 down, 400 up on Speedtest to Spark Wellington.. and I'm presuming here that the Spark Wellington Speedtest.net servers are actually in Wellington.

Ping between the 2 Palmerston North connections is 13 ms... router to router.

I'm trying to understand or have someone in the know explain to me how all this works.

The goal is low latency between the two same city sites (1-2 ms would be nice :-)) as we are running an SD-WAN solution between the two.

Any insider insight into this would be appreciated.

1 | 2

30853 posts

Uber Geek
+1 received by user: 9996

Retired Mod

Trusted

Biddle Corp

Lifetime subscriber

#2712409 24-May-2021 16:10

Most obvious answer would be that Spark don't actually have a BRAS in Palmy, and connections actually terminate somewhere else.

yitz

2238 posts

Uber Geek
+1 received by user: 594

#2712410 24-May-2021 16:12

Could be a Static IP thing and related to how traffic is routed?

nbroad

320 posts

Ultimate Geek
+1 received by user: 39

#2712425 24-May-2021 16:27

Yes that's the question really.. do Spark have the infrastructure (what is it called? a BNG, BRAS, handover or router?) in Palmerston North to route traffic locally between connected UFB customers and how do I get on it!

Is it a Chorus provisioning thing?

Does requesting a static IP make a difference?

Handover mapping - is that what I'm looking for?

Cheers

yitz

2238 posts

Uber Geek
+1 received by user: 594

#2712438 24-May-2021 16:51

If you want to test whether Static IP has an effect you could temporarily switch from static to dynamic by assigning PPP username as NoStatic (any password) on both ends.

Also, I recall BNG name is contained in the PPPoE Access Concentrator field which you router may show on its web UI or CLI.

sparkz25

750 posts

Ultimate Geek
+1 received by user: 284
Inactive user

#2712444 24-May-2021 17:05

What about Direct fiber access?

or are you close enough to set up a wireless link?

nbroad

320 posts

Ultimate Geek
+1 received by user: 39

#2712445 24-May-2021 17:11

sparkz25:

What about Direct fiber access?

or are you close enough to set up a wireless link?

We are just moving away from Spark WAN Services to SD-WAN to save some $$$.

I do actually have an AirFiber wireless link up, but I use that for backup and SAN replication traffic.

Just wanting to see if I can get the UFB path(s) configured as optimal as possible.

Cheers

Geekzone

Want to support Geekzone and browse the site without the ads? Subscribe to Geekzone now (monthly, annual and lifetime options).

nztim

4012 posts

Uber Geek
+1 received by user: 2710

ID Verified

Trusted

TEAMnetwork

Subscriber

#2712505 24-May-2021 17:56

Unless the UFB provider you use has a BNG in Palmerston North all traffic will be routed to the nearest BNG (Most likely Wellington or Auckland)

this is definitely a case for DFAS which is circa $400 each tail plus a tie cable service

You may also need ICABS if the two sites fibre terminate at different exchanges

Any views expressed on these forums are my own and don't necessarily reflect those of my employer.

cyril7

9073 posts

Uber Geek
+1 received by user: 2499

ID Verified

Trusted

Subscriber

#2712575 24-May-2021 18:50

Inspire offer excellent local traffic options and local private WAN to boot.

That said from an Inspire site in the city to a voda BIA connection also in the city its 75mS via Syd, but to a voda Vdsl at the same location is 20mS via AKL, go figure

Cyril

chevrolux

4962 posts

Uber Geek
+1 received by user: 2638
Inactive user

#2712596 24-May-2021 19:55

Yea get some Inspire Net fibre in mate - I thought you already would have to be honest! You'll easily see your 1-2ms between sites.
Even their standard UFB products (like basic GPON ufb) handover over in town and latency is super low between sites.

Spark defibitely have a handover in town, but I guess they haul it off down to Wellington.

hio77

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#2712599 24-May-2021 20:03

have you looked at a traceroute to see where the latency jump actually is?

it sounds like both of your connections are terminating in wellington.

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

boosacnoodle

1268 posts

Uber Geek
+1 received by user: 855

#2712610 24-May-2021 20:40

What is the application that 13ms is likely to make a difference?

Lenovo

Shop now for Lenovo laptops and other devices (affiliate link).

Kraven

738 posts

Ultimate Geek
+1 received by user: 190

#2712611 24-May-2021 20:46

I get 1-2ms between my BigPipe UFB connection and a couple of customer Spark UFB connections in Palmy. The reported BNG on my BigPipe connection is PM-BNG02.

As others have suggested though, shift to Inspire if you don't want traffic between sites leaving the city.

nbroad

320 posts

Ultimate Geek
+1 received by user: 39

#2712613 24-May-2021 20:52

boosacnoodle:
What is the application that 13ms is likely to make a difference?

So far the only complaint I have had from users is a report with a poorly written SQL query taking a fair bit longer to run, so it’s not the end of the world.

nbroad

320 posts

Ultimate Geek
+1 received by user: 39

#2712630 24-May-2021 20:55

Kraven:
I get 1-2ms between my BigPipe UFB connection and a couple of customer Spark UFB connections in Palmy. The reported BNG on my BigPipe connection is PM-BNG02.

As others have suggested though, shift to Inspire if you don't want traffic between sites leaving the city.

Good to know that technically it is possible, thanks for that.

hio77

'That VDSL Cat'
13036 posts

Uber Geek
+1 received by user: 3896

ID Verified

Trusted

Lizard Networks

Subscriber

#2712659 24-May-2021 21:28

boosacnoodle:
What is the application that 13ms is likely to make a difference?

Sql, storage, clustered applications can all be pretty latency sensitive.

Sql is generally an easy one as the Soultion is to write the application to make bulk requests rather than synchronous. But not always an option.

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

1 | 2