Premium vs standard transit to China

How premium China routing differs from standard international transit: what changes, what stays the same, and which questions decide the fit.

Premium China routing and standard international transit differ in path, capacity and responsibility, not just speed. Premium uses a carrier-backed route for internal traffic; standard transit uses the public internet's path mix. The right fit depends on where the application runs and who it serves.

What premium routing changes

Premium China routing is a carrier-backed path for your company's internal traffic. Standard international transit uses the public internet path mix. Filing and data-transfer steps stay with you and your counsel in both cases. The difference is not a faster pipe for every packet; it is a different path and a different capacity model.

Where the application serves visitors inside China, the filing and hosting questions dominate the decision; where it serves your own sites and devices, the traffic profile does. Start there, then compare the technical options. A premium path does not remove a filing obligation, and a standard path does not create one where none applies.

Two paths, two responsibility models

What changes between a premium China route and standard international transit, and what the provider owns in each case.
ItemPremium China routeStandard international transit
PathCarrier-backed, engineered for China-bound trafficPublic internet path mix, chosen by each network
CapacityCommitted rate plus burst, stated in the contractBest effort; the port rate is not a reservation
ResponsibilityFiling and data transfer remain with you; ask what support the quote includesProvider sells transit; filing remains with you
FitInternal sites, plants, offices and device backhaulPublic content, CDN delivery and general reach

Neither path is a compliance determination. Both carry your traffic only; neither is a public internet service for third parties.

MIIT Circular [2017] No. 32 via CAC; our product definition

The fiber floor applies to both paths

40 ms

The fiber floor applies to both paths
Value What it means
10 ms Regional path
50 ms Intercontinental path
160 ms Queues or a long path
300 ms Queues or a long path
800 ms Queues or a long path
Reference marks for round-trip time. They are not measurements of our network; the tool computes the floor for a city pair. Source: ITU-T G.114 and Y.1541: transmission time and IP performance objectives
Text description

The scale marks 10, 50, 160, 300 and 800 milliseconds. A same-metro path sits at the low end. A regional path is tens of milliseconds. An intercontinental path starts at the fiber floor, about 1 ms round trip per 100 km of one-way distance; a trans-Pacific fiber round trip of about 16,000 km has a floor near 160 ms before detours, and queues or a longer path push the figure higher. The marks are reference points, not a claim about any provider.

Ask for the same fields in both quotes

A quote comparison template. No field here is a price.
FieldPremium quoteStandard quote
PathNamed route, carrier-backed, internal trafficPublic internet path mix
CapacityCommitted rate plus burst, stated in the contractBest effort; port rate is not a reservation
ResponsibilityFiling and data transfer remain with you; ask what support the quote includesProvider sells transit; filing remains with you
MeasurementProbe locations, window, packet size, one-way or round-tripSame fields; one speed test is not a baseline
HandoffLayer 2 or Layer 3, named endpointsLayer 3, upstream BGP
ExclusionsEquipment, installation, testing, moves and changesSame list; compare the excluded scope

A low monthly line item with a large excluded scope is not a lower-cost option. Compare the scope before the number.

Our quote-review template; service definitions: ITU-T G.8011

Separate physics from routing

The tool uses distance and a path factor. It does not measure a live route.

The physical minimum round-trip time between two cities over fiber. Physics only, not our network.

pathFactor
1.5
Open the tool

Latency: what a premium path can and cannot fix

Light in fiber travels at roughly 200,000 km per second, about two thirds of the speed of light in a vacuum. That sets a physical floor: a 10,000 km path has a one-way propagation time of about 50 ms before any router, queue or detour. A premium path can remove a long detour and reduce jitter by keeping traffic on a controlled route. It cannot beat the fiber floor, and it cannot fix an application that makes six sequential round trips.

Measure round-trip time, jitter and loss from both ends, at peak and off-peak. Use the fiber latency floor tool to separate physics from routing. If the measured round-trip time is close to the floor, the remaining problem is usually application design or congestion, not the path.

Loss, jitter and congestion

Packet loss and jitter are the usual reasons a China-facing application feels slow even when the latency looks acceptable. They come from congestion, route changes, wireless links and overloaded equipment. Premium capacity is usually sold with a committed information rate and a burst allowance; standard transit is best effort. Ask what is committed, what is measured, over what window, and what happens when the measurement misses. A number without a method is not a service definition.

ITU-T Y.1541 defines IP quality-of-service classes and objectives, and ITU-T G.114 gives guidance on one-way transmission time for voice. Use them to frame the questions, not as a substitute for a route test.

Security and exposure

A private line is private in the sense that it is not on the public internet; it is not encrypted by default. Run IPsec or TLS over it as you would over any other path, and keep the same identity and access controls. Standard transit exposes the service to the public internet, so DDoS filtering, a web application firewall, rate limits and a tested failover matter more than the path choice. A private path can still be reached through a public-facing application. Ask where the handoff sits, which side owns the router, and whether the service includes any filtering. Do not assume that a private path removes the need for security controls.

Compliance and filing

MIIT Circular [2017] No. 32, issued by MIIT and reproduced by the Cyberspace Administration of China, targets unapproved cross-border dedicated lines used to conduct cross-border business. It also requires basic telecom operators that lease international private lines to keep user records and to state that the lines are for internal office use only.

Where a website is hosted inside mainland China, an ICP filing is made for each domain through the hosting or CDN provider. MLPS 2.0 grades systems operated in China and sets the controls for each grade. Personal information leaving China follows the Personal Information Protection Law and its export mechanisms. Neither premium routing nor standard transit removes these steps; scope them with your counsel. This is general information, not legal advice. Regulations change, so check the official texts and your own obligations. The compliance primer covers the rule and the steps in more detail; the ICP filing and data residency entries cover the vocabulary.

How to compare two quotes

Ask for the route in writing: the handoff type (Layer 2 or Layer 3), the endpoints, the last mile, the committed rate and the burst. Ask which party provides the router, the cross-connect and the local loop. Ask for the measurement plan: probe locations, test window, packet size, and whether the result is one-way or round-trip. Ask how a fault is escalated and who owns the carrier interface on each side. A support model describes how faults are handled; it does not state an outcome. Ask what is excluded: filing support, equipment, installation, testing, moves and changes. The lowest monthly line item often has the largest excluded scope.

Which option fits

Internal sites, factories, offices and device backhaul usually fit a premium private path because the traffic is the company's own and the path can be controlled. Public visitor traffic is different: filing and hosting dominate, and a CDN plus standard transit may be enough. A hybrid is common: public content on standard transit or a CDN, and a private path for internal systems, management access and device data. Keep the two roles separate in the design so a filing or security issue on one does not force a redesign of the other.

Where the application serves visitors inside China, do not buy a private line and assume the filing question disappears. Where it serves only your own sites, do not buy public transit and assume the internal-use question disappears. Write the use case down in one sentence before the quote: who the users are, where the systems sit, and whether the traffic is internal or public. That sentence is the scope the rest of the comparison hangs on. For the product vocabulary, see IP transit, IEPL vs IPLC, China routing explained, and how networks are named in routing data.

Common mistakes

Comparing Mbps only. Ignoring the last mile. Assuming premium fixes application latency. Assuming standard transit is unregulated. Forgetting that a private line is not encrypted. Buying a route without a baseline measurement. Treating a quote as a service definition. Leaving the filing question to the network team alone.

What stays the same

Both options carry your traffic only; neither is a public internet service for third parties. Both need the same commercial basics: a signed contract, a defined traffic scope, and contacts on both ends. Both also need a named owner on your side for changes, incidents and billing. This is not legal advice. The product definition is scoped per quote; ask for a quote and we scope the current options with you.

Before you choose

This list is a buying aid, not legal advice.

Check what you learned

  1. Premium routing is not a faster pipe for every packet. What is it?

    A different path and capacity model: a carrier-backed route for internal traffic. Filing and data transfer still need scoping with your counsel. The fiber floor still applies.

  2. Both quotes show the same round-trip time. Does that settle the comparison?

    No. Compare the path, the committed rate, the measurement method, the exclusions and who owns the handoff. The same average can hide different loss and jitter.

  3. The application serves visitors inside China. Does a private line remove the ICP filing question?

    No. A private line changes the path. ICP filing depends on hosting and on serving users inside mainland China. Scope the application with counsel.

Answers are revealed on request. Nothing is scored, stored or sent.

FAQ

Which option do visitors to China need?

That depends on filing, hosting and traffic shape more than on bandwidth. Describe the use case in a quote request and we scope it with you.

Is premium routing always faster?

No. It can remove a detour and reduce jitter, but physics sets a floor and application design can dominate. Measure both paths from both ends.

Do we need ICP filing with standard transit?

ICP filing is generally tied to hosting inside mainland China, not to the transit product. A site served from outside mainland China is usually outside the filing system; a site hosted inside needs a filing for each domain. Confirm your case with counsel.

Is a private line encrypted?

No, not by default. Private describes the path, not the encryption. Use IPsec or TLS and keep the same access controls.

What should we measure before choosing?

Round-trip time, jitter, loss and time to first byte, from both ends, at peak and off-peak, over a window long enough to include a busy period.

Is this legal advice?

No. This page is general information. Check the official texts with your counsel.
  • Entering China compliance checklist

    The checklist for China-bound connectivity: filing, data-transfer scoping, and the single contact and escalation path. Not legal advice.

  • RFQ and RFP template pack

    Field lists for SMS wholesale, eSIM and data, IoT and China connectivity requests, plus the online checklist that sends as an RFQ.

  • AS number

    An autonomous system number identifies the network that announces an address prefix, and the AS path shows which networks the route crosses.

Adjacent to what we offer

We explain this here because it affects how you use our service.