[Please note that the information around both the NBN and UFB initiatives is still quite fluid. Apologies in advance for any inaccuracies in the below text – Ed].
An important first point here is that I am not discussing the Rural Broadband Initiative in NZ, so I will stay away from mentioning wireless and satellite services aimed at remote Australian areas.
We will just stick to layer fibre.
Secondly, the approaches are fundamentally different in that NBN co is a single entity rolling out fibre nationwide and then wholesaling it to ISPs. However in NZ, four Local Fibre Companies (LFC’s) were signed by Crown Fibre to roll out services in selected geographical areas, resulting in nuanced differences across regions.
Circuit Access Speeds
New Zealand
|
Australia
|
30Mbps/10Mbps (GPON) |
12Mbps/1Mbps |
100Mbps/50Mbps (GPON) |
25Mbps/5Mbps |
100Mbps (GPON/P2P) |
25Mbps/10Mbps |
1Gbps (P2P) |
50Mbps/20Mbps |
10Gbps (P2P) |
100Mbps/40Mbps |
It seems to me that NZ has the edge on all fronts here, with a faster entry level service, more symmetrical options, and a faster maximum speed range. I am not sure why the NBN has chosen to go with asymmetrical speeds across the board, or why they have tapped out before surpassing 100Mbps download. Business needs are the same across both countries, so I don’t see how demand can be met through the Australian offering, but it does seem to suggest that existing networks have a substantial market space to keep a share of in that market. Australia is rolling out GPON in isolation, whereas the UFB rollout comprises a P2P suite of services and dark fibre in addition to the cheaper GPON tails, making it a far more complete solution for users in every demographic.
Another difference between the approaches of our two countries is that the NBN does not distinguish, whether it is based on price or speed, between residential or business access. NBN instead relies on traffic classes and ISP contention levels to distinguish between cost and performance.
Australian Traffic Classes
Traffic
Class |
Description |
Notes |
1
|
Designed for Telephony
|
Available now |
2
|
Designed for high priority video traffic
|
Intended for future release – no ETA set
|
3
|
Designed for generic high priority data
|
Intended for future release – no ETA set
|
4
|
Best effort only |
Available now |
I have been unable to find any specific metrics on jitter, latency, or packet loss for the above classes, but if anyone else has been successful in this area I would love to hear from you.
New Zealand Traffic Classes
Traffic
Class |
Frame Delay |
Frame Delay Variation |
Frame Loss |
CIR |
≤ 5ms |
≤ 1ms |
≤ 0.1% |
EIR |
N/A |
N/A |
≤ 2% |
NZ has a less complex solution in this space, and although you can tailor your QoS structure to a greater degree in Australia, or rather, you will be able to once the additional classes are released into production, the less complex a solution is, the easier it is to succinctly package and take to market.
RSP hand-off design
To avoid confusion here I will again just focus on layer 2 services so we have symmetry between each country’s product sets.
NZ has been split up into 33 regions and an RSP can connect in a couple of ways. They can deploy a rack in an LFC’s exchange, or they can purchase an intraregional backhaul service to terminate services in an existing POP. There are a few issues with this, namely, that both processes are very expensive in terms of recurring costs and insurance overhead when deployed nationwide, and at present Chorus does not have an inter-regional backhaul service.
NBN works in a similar manner except that it has a very clear focus on providing national backhaul through its “connectivity virtual circuit” where an ISP can purchase traffic classes on a per Mbps basis and get end user traffic back to their core POP’s.
Both scenarios have the same issue in that only a select few ISP’s have the resources to connect to either network nationwide and as a result both markets are seeing the rise of third party wholesale aggregators, entities which will provide services in low yield regions to ISP’s that can only justify rolling out UFB/NBN interconnects in key metro markets.
SLA’s
The NZ UFB rollout is far more advanced than its Australian counterpart, and consequently the information available around fault restoration is almost nonexistent at this time for the NBN. What we do know is that the fibre network has a target of 99.90 per cent uptime and that the highest level traffic class (TC-1) has the following service parameters:
- Frame delay variation ≤ 16ms
- Frame Delay ≤ 25ms
- Frame Loss ≤ 0.04%
As the premier traffic class, these metrics sit well below the standards for CIR in the UFB space, although having said that, the frame loss target is tighter or TC-1.
The lowest traffic class is simply best effort, and the other two options haven’t been released as yet.
Chorus has the worst of the NZ LFC’s SLA’s, but none of them are good.
Crown Fibre missed a chance to justify the premium on the P2P products by attaching a faster restoration target that included faults on the termination device (ONT), but as it stands, you will wait for up to 48 hours for service restoration on all services unless you pay extra, either monthly, or as a one off $700 charge, and of the three SLA options provided by Chorus, only the highest level (and therefore the most expensive) includes faults to the ONT.
If I am not off base, then as the premier traffic class, it is sits well below the standards for CIR in the UFB space, although the frame loss is a tighter metric within the NBN.
Pricing
As previously mentioned, the NBN does not distinguish between residential and business last mile access. As such we see the following price points:
Australia |
Wholesale Price
|
12Mbps/1Mbps |
$24
|
25Mbps/5Mbps |
$27
|
25Mbps/10Mbps |
$30
|
50Mbps/20Mbps |
$34
|
100Mbps/40Mbps |
$38
|
The above numbers are attractive, but are not indicative of the final cost. All traffic classes used for backhaul to the ISP will cost $20 p/Mbps, so contention ratios will dictate retail price points given it is this portion of service delivery that attracts the greatest overheads. On top of backhaul, co-location, and IP transit will combine to determine final retail price points.
The UFB rates start at $37.50 for a residential 30Mbps/ 10Mbps (2.5Mbps each way) GPON circuit, then goes up to $480 for a 100Mbps CIR symmetrical P2P circuit (and beyond). But again, these prices are not indicative of final price points.
At present DTS, Orcon and iiNET are a few of a small number of ISP’s that have publicly stated UFB/NBN retail price points on their websites.
Summary
The UFB network is designed to provide a connectivity up-grade nationally, and where fast metro networks are already in place, it seeks to encourage adaptation to ensure attractive price points through more cost effective delivery.
Conversely, The NBN is looking to bring connectivity where there essentially is none at present, or it is of such poor quality that modern applications are not able to be considered for use. Where it is rolled out in metro areas, the residential market and small business segments will be keenly interested, but larger businesses will likely be better served by existing network providers.
Through offering more service layers and multiple delivery methods, the UFB product suite is scalable and future focused, whereas the NBN service platform will need to be substantially upgraded to have anything close to a comparable benefit to businesses.
Brendan Ritchie is the Australian-based CEO of DTS, a business-focused ISP operating in both Australia and NZ.
Twitter profiles: @DTSnz and @DTSaus
GLOSSARY OF ACRONYMS
- P2P — Point to point
- GPON — Gigabit Passive Optical Network
- RSP — Registry Service Provider
- UFB — Ultra Fast Broadband, the NZ government’s fibre to the home initiative
- NBN — Australia’s high speed broadband network
- POP — Post Office Protocol
- CIR — Committed information rate
|