What Is IP Transit? A Complete Guide to How It Works, Pricing, and Providers

Quick Answer
‍
IP transit is a paid service in which an upstream network carries your traffic to and from every reachable destination on the internet. The provider advertises your IP prefixes globally over BGP and hands your routers the full internet routing table. You typically pay for a committed bandwidth level on a physical port, billed at the 95th percentile of measured throughput.

What Is IP Transit?

IP transit is a commercial arrangement in which one network pays another to carry its traffic to the rest of the internet. The relationship is hierarchical rather than mutual. Your provider sits upstream, accepts your route advertisements, propagates them to its own upstreams and peers, and accepts return traffic on your behalf. That single relationship is what makes an autonomous system globally reachable.

The distinction that matters commercially is scope. Transit sells reachability to every network on the internet, including ones the provider has no direct relationship with. A peer reaches only itself and the networks behind it. That is why a network can hold hundreds of peering sessions and still be unreachable from parts of the internet without transit.

Most enterprises buy transit implicitly inside a hosting or colocation contract. Content platforms, gaming operators, CDNs, and anyone running their own ASN buy it explicitly as carrier-grade IP Transit on a dedicated port, with routing policy under the buyer’s control.

Not sure whether you need dedicated transit or blended connectivity inside a contract? 

Talk to a NetActuate Engineer.

How Does IP Transit Work?

IP transit runs over BGP sessions between your routers and your provider’s. The provisioning sequence is consistent across the industry.

1. Physical connection. A cross connect or port is provisioned between your equipment and the provider, commonly at 1G, 10G, 40G, 100G, or 400G. (NetActuate offers 1G to 100G)
2. BGP session establishment. An eBGP session is configured between your ASN and the provider’s over the assigned point to point subnet.
3. Prefix advertisement. You advertise your prefixes. The provider validates them against your Letter of Authorization and IRR or RPKI records, then propagates them upstream.
4. Full table delivery. The provider sends the complete internet routing table, now past one million IPv4 routes, so your routers can pick a next hop for any destination.
5. Traffic forwarding and return. Outbound leaves through the provider. Inbound follows the paths your advertisements created, which is why prepending and communities shift inbound load.

Route propagation is not instantaneous. A new advertisement converges globally within minutes, and a withdrawal can take longer to clear from distant tables. That window, not port capacity, determines how fast a failed uplink stops receiving traffic. Networks needing faster failover pair transit with BGP Anycast routing so multiple locations announce the same prefix simultaneously.

What the Full Routing Table Actually Buys You

The full table is what separates transit from a default route (NetActuate IP Transit can also be delivered as default-only or filtered). With a default route your router sends anything unknown to one upstream and accepts whatever path it chooses. With the full table it weighs AS path length, local preference, and communities per destination and decides for itself. That control is why multihomed networks buy transit rather than plain internet access, and you can check how any provider sees a prefix using a public looking glass.

IP Transit vs Peering: Where Your Traffic Actually Exits

Most buying guides treat transit and peering as a choice. In production you use both, and the ratio between them determines delivery quality and effective cost. Every quote is a price per megabit of committed capacity. It does not tell you what share of your traffic leaves through settlement free peering versus being handed to a paid upstream who hands it onward again.

That share is the blend ratio, and it is the most useful question to ask a provider. A thin network with two upstreams and a short peering list hands almost everything to a paid upstream, adding AS hops and putting latency at the mercy of someone else’s congestion policy. A densely peered network delivers most traffic straight to the destination ASN. The unit price can be identical while delivered performance is not.

This is why direct IX peering complements rather than replaces transit. Peering at internet exchange points removes cost and AS hops for destinations on that fabric, while transit covers everything else. For the mutual exchange side, see our explainer on how network peering works.

NetActuate delivers blended transit across one of the world's most densely peered networks, with peering and Tier 1 routes on the same port.

Request an IP Transit Quote.

How Much Does IP Transit Cost?

IP transit is commonly quoted per megabit per second per month against a committed rate (NetActuate offers the choice between per-GB, per-Mbps, flat-rate port, 95th-percentile burstable, or aggregate volume commitments across regions). Published TeleGeography research shows sustained price erosion, with 100 GigE port prices in major hubs falling at a double digit compound annual rate and wide regional spreads persisting. Northern Virginia, London, Amsterdam, and Frankfurt sit at the bottom of the range. Regions with limited subsea diversity price several times higher for identical capacity.

Treat the per megabit figure as the last question, not the first. Route quality, commit structure, and upgrade flexibility move annual spend more than the headline rate.

How 95th Percentile Billing Works

Your provider polls the interface every five minutes, producing roughly 8,640 samples per month. The top five percent are discarded and the highest remaining sample sets your billed rate. That discard is worth about 36 hours, which is why a nightly backup window rarely moves the invoice but sustained daytime growth does.

Commit to Port Ratio: The Cost Lever Most Buyers Miss

Your committed rate and your port size are separate purchases. A common structure sets the commit near a third to a half of port capacity, leaving headroom to burst without renegotiating. Overcommitting to chase a lower unit rate is the most frequent budgeting error here, because the discount only pays off if the traffic arrives. Ask for the fully loaded monthly figure including port, cross connect, and setup before comparing quotes, and confirm dual stack with full BGP carries no surcharge.

Want your current transit invoice modeled against a blended alternative before you renew?

‍How to Evaluate IP Transit Providers

Evaluate the network, not the brochure. These criteria separate otherwise comparable quotes.

• Peering density and blend ratio. Ask what share of traffic to your top destination ASNs exits via peering rather than paid upstream, per region.
• Path evidence, not claims. Run traceroutes from the provider’s looking glass to your real user networks and compare AS path length against your incumbent.
• Regional footprint alignment. Transit bought where your users are outperforms transit bought where it is cheapest. Confirm capacity in your real delivery markets.
• Congestion and capacity policy. Ask at what utilization the provider augments a peering link or upstream, and what the historical lead time has been.
• Integrated protection. Volumetric attacks arrive over the transit port. Confirm whether DDoS mitigation is applied upstream of it or sold separately.
• Contract flexibility. Check mid contract upgrade terms, whether commits move between locations, and what burst above commit costs.

NetActuate operates blended transit across 50+ global points of presence, combining Tier 1 upstreams with a large peering base on the same port, with IPv4, IPv6, and BGP support included. Transit deploys standalone to third party infrastructure or alongside bare metal, cloud routers, and the wider networking suite.

Learn more about NetActuate IP Transit.

Frequently Asked Questions About IP Transit

What is IP transit in simple terms?

IP transit is a paid service in which an upstream network carries your traffic to and from every destination on the internet. The provider advertises your prefixes globally over BGP and hands you the full routing table.

What is the difference between IP transit and peering?

Transit is a paid, hierarchical relationship that delivers your traffic to the entire internet. Peering is a direct exchange between two networks, usually settlement free, reaching only the networks behind each peer.

How much does IP transit cost per Mbps?

Pricing depends on region, port size, and commit level, and the spread between markets is large. Dense hubs sit at the bottom of the range while regions with limited subsea capacity price materially higher.

What is 95th percentile billing?

Your provider samples throughput every five minutes, discards the highest five percent of samples, and bills the highest remaining one. That discard covers roughly 36 hours of burst, so short spikes rarely change the invoice.

How many IP transit providers should a network use?

Two upstreams is the practical minimum, because a single transit relationship makes one provider a point of failure for global reachability. Blended transit from a densely peered provider often reaches comparable diversity with fewer contracts