Internet Exchange Points: Guide to IXPs and Network Peering

Internet exchange points are physical locations where multiple networks connect to exchange traffic directly with each other. Instead of routing traffic through expensive transit providers who charge per megabit, networks at IXPs peer directly – often at zero cost beyond the port fees to connect to the exchange. This direct exchange reduces latency, cuts bandwidth costs dramatically, and improves performance for end users.

Without IXPs, every network would need to pay transit providers to reach every other network. A content delivery network serving video to residential ISPs would pay transit charges for every gigabit transferred. With IXPs, that same CDN connects to the exchange point, peers directly with the ISPs, and transfers terabits at zero marginal cost. The savings are substantial enough that IXP connectivity becomes mandatory for any organization moving serious traffic volumes.

Most internet users have no idea IXPs exist, but nearly every piece of internet traffic flows through at least one exchange point. When you stream video, load websites, or use cloud applications, that traffic is probably traversing an IXP somewhere along the path. These exchanges form the glue that holds the internet together, enabling the massive traffic volumes that modern applications generate.

What Are Internet Exchange Points?

Internet exchange points are neutral facilities where network operators connect their infrastructure to exchange traffic with other networks. The IXP provides switching infrastructure – typically Ethernet switches – that all participating networks connect to. Once connected, networks establish bilateral peering agreements with each other and begin exchanging traffic directly.

The “exchange” in internet exchange point refers to traffic exchange between networks. When your ISP’s network needs to deliver traffic to Netflix, that traffic exchanges from one network to the other. If both networks participate in the same IXP, that exchange happens locally at very low cost rather than through expensive intermediary networks.

How IXPs Function Technically

At the physical layer, an IXP operates a switch fabric – interconnected Ethernet switches that provide connectivity between all participants. Each network operator installs equipment in the IXP facility and connects to the exchange’s switch fabric through one or more ports.

Port speeds at major IXPs range from 1 Gbps for smaller participants to 400 Gbps for large networks moving massive traffic volumes. The port connects your router to the IXP’s switch fabric, giving you Layer 2 connectivity to all other IXP participants.

Once physically connected, networks establish BGP peering sessions with each other to exchange routing information. These peering sessions advertise which IP prefixes each network can reach, enabling traffic to flow directly between networks rather than through transit providers.

The IXP itself doesn’t make decisions about who peers with whom. Networks independently decide which other networks they want to peer with based on their own policies, traffic ratios, and business relationships. The IXP just provides the physical infrastructure and switching fabric that makes these connections possible.

Who Operates IXPs

IXPs are typically operated as neutral, non-profit organizations or cooperatives owned by participating networks. This neutral model ensures the exchange doesn’t favor any particular network and operates purely to facilitate interconnection.

Some IXPs operate commercially but maintain neutrality through transparent policies and open participation requirements. The key is that the exchange operator doesn’t compete with networks using the exchange and doesn’t give preferential treatment to specific participants.

Regional exchanges often start through grassroots efforts by local network operators who recognize the value of local traffic exchange. As the exchange grows and proves valuable, more networks join, creating network effects where each new participant makes the exchange more attractive to others.

IXPs vs. Private Interconnection

IXPs provide public peering, where you connect to a shared switch fabric and establish peering with multiple networks through that infrastructure. This differs from private interconnection, where two networks establish a dedicated cross-connect between their equipment.

Public peering through IXPs scales better when you need to reach many networks. One port to the IXP potentially connects you to dozens or hundreds of other networks. Private interconnection makes sense for high-volume bilateral relationships where dedicated capacity justifies the cost.

Many organizations use both. Establish private interconnections with your highest-volume peers where dedicated capacity makes sense. Use IXP peering to reach the long tail of smaller networks where individual traffic volumes don’t justify private connections.

How Internet Exchange Points Improve Network Performance

The performance benefits of IXP connectivity come from several factors that make direct exchange more efficient than routing through transit providers.

Latency Reduction

Traffic flowing through IXPs takes fewer hops than traffic routed through transit providers. Every router hop adds latency – typically 1-2 milliseconds per hop. Transit routing might send traffic through 10-15 hops to reach the destination network. IXP routing reduces this to 2-3 hops.

For latency-sensitive applications, this difference matters. A video conferencing call where both endpoints’ ISPs peer at an IXP might see 15-20 milliseconds lower latency than if traffic routes through transit providers. Gaming, real-time communications, and financial applications all benefit from these latency reductions.

The latency advantage compounds when multiple parties are involved. A multi-party video conference with participants on different networks benefits when all those networks peer at common IXPs rather than routing through diverse transit paths.

Path Optimization

Transit routing often takes geographically inefficient paths because routing decisions prioritize business relationships over geography. Traffic between two networks in the same city might route through a transit provider’s backbone in another state, adding hundreds of miles of unnecessary distance.

IXPs provide local traffic exchange. When both networks connect to the same IXP in the same city, traffic takes the direct local path. This optimization is why most major metro areas host IXPs – local networks recognize the value of local traffic exchange.

Content delivery networks particularly benefit from IXP presence in many markets. They cache content at locations around the world and peer with local ISPs at regional IXPs. This keeps video streaming and content delivery traffic local rather than routing through backbone networks.

Packet Loss Reduction

Transit paths often experience congestion during peak hours, leading to packet loss when queues overflow. IXP paths typically see less congestion because they’re not sharing capacity with transit traffic to diverse destinations.

Lower packet loss improves application performance. TCP throughput degrades with packet loss, so even small improvements in loss rates translate to better user experience. UDP-based applications like video streaming and voice calls handle packet loss better than TCP, but still benefit from cleaner paths.

Traffic Engineering Control

When you peer at IXPs, you gain more control over routing decisions. You can prefer IXP paths over transit paths through BGP communities and local preference settings. This lets you optimize costs by routing as much traffic as possible through zero-cost peering rather than expensive transit.

Traffic engineering through IXPs also provides better redundancy. If your primary transit provider has issues, IXP peering provides alternative paths that keep your network reachable. This redundancy improves overall reliability compared to depending entirely on transit providers.

Benefits of Direct Peering vs. Transit

Understanding the difference between peering and transit clarifies why IXP connectivity delivers value.

What is Transit?

Transit means you pay a network provider to carry your traffic to the rest of the internet. The transit provider accepts all your traffic and uses their connections to other networks to deliver it to destinations. You pay per megabit of capacity, and those charges accumulate monthly based on your peak or average usage.

Transit makes sense when you’re starting out or when you need to reach networks you can’t peer with directly. Every network uses some amount of transit. The question is how much you can offset through peering to reduce costs.

Transit pricing varies by market and provider in major markets. Smaller markets or providers with less competition charge more. At scale, these costs add up quickly.

What is Peering?

Peering means two networks agree to exchange traffic directly without payment in either direction. This “settlement-free” peering is the default model at IXPs. Both networks benefit from improved performance and reduced transit costs, so neither charges the other.

Some peering relationships involve payment when traffic ratios are very unbalanced, but IXP peering is typically settlement-free. The only costs are the port fees to connect to the IXP.

The peering approach saves over $1 million annually. Even accounting for equipment costs and engineering time, the ROI is obvious. This is why content providers, CDNs, and large enterprises all pursue IXP peering aggressively.

Performance Benefits

Beyond cost, peering improves performance through lower latency and fewer hops. Transit routing can’t match the direct paths peering provides. For user-facing applications, this performance improvement translates directly to a better experience and potentially better business outcomes.

Content providers streaming video find that lower latency and better routing through peering reduces buffering and improves video quality. This keeps users engaged and reduces churn compared to poor streaming experiences through congested transit paths.

Public vs. Private Peering Options

Peering happens in two main configurations – public peering through the IXP switch fabric and private peering through dedicated connections.

Public Peering Explained

Public peering connects networks through shared IXP infrastructure. You connect to the IXP’s switch fabric and establish peering sessions with other participants through that shared infrastructure. All your peering traffic with all your peers at that IXP flows through your single port.

This model scales efficiently. One 100 Gbps port to an IXP can carry traffic to dozens or hundreds of peers. As you add new peering relationships, you use more of your port capacity but don’t need additional physical connections.

Public peering is “public” only in the sense that you peer through shared infrastructure. Your peering relationships and traffic volumes remain private. Other networks at the IXP can see that you participate but can’t see your specific traffic flows or peering arrangements.

Private Peering Explained

Private peering establishes dedicated physical connections between two networks, usually through cross-connects in data centers with carrier-neutral interconnection. Instead of routing through the IXP switch fabric, traffic flows over a dedicated link between your equipment and your peer’s equipment.

This approach makes sense for high-volume bilateral relationships where traffic justifies dedicated capacity. If you’re exchanging 40 Gbps with a specific peer, a dedicated 100 Gbps private interconnect provides ample capacity and eliminates concerns about IXP port congestion.

Private peering also provides better security since traffic doesn’t traverse shared switch fabric. Some organizations with sensitive traffic prefer private interconnection even when traffic volumes might not strictly justify it from a capacity perspective.

When to Use Each Approach

Use public peering when:

• You want to reach many networks without establishing individual connections to each one
• Your traffic with individual peers is modest relative to your IXP port capacity
• Cost efficiency matters more than dedicated capacity
• You’re establishing peering relationships opportunistically rather than negotiating large bilateral agreements

Use private peering when:

• You exchange massive traffic with specific peers (10+ Gbps per peer)
• You need guaranteed capacity for specific peering relationships
• Security or compliance requirements favor dedicated connections
• Your peer is in the same facility, and cross-connects are readily available

Many organizations implement hybrid approaches. Use public peering at IXPs to reach numerous networks efficiently. Establish private peering with your highest-volume peers where dedicated connections make sense. This balances cost efficiency with performance for high-volume relationships.

Transitioning Between Models

Peering relationships often start as public peering through IXPs. As traffic grows, networks upgrade their IXP ports to handle increasing volume. Eventually, traffic with specific peers might grow large enough that private peering makes sense.

This transition is normal and reflects successful peering relationships. Starting with public peering keeps initial costs low while you prove out the relationship. Moving to private peering later optimizes costs and performance for high-volume partnerships.

Internet Exchange Points Participation Requirements

Joining an IXP requires meeting specific technical and policy requirements. Understanding these requirements helps you prepare for IXP participation.

Technical Requirements

You need your own Autonomous System Number (ASN) and provider-independent IP address space. These are prerequisites for BGP peering with other networks. Without your own ASN, you can’t participate in IXPs as an independent network.

Obtaining an ASN and IP space involves working with Regional Internet Registries (RIRs) – ARIN for North America, RIPE for Europe, APNIC for Asia Pacific, and others. The process requires justifying your need for resources and paying annual fees to maintain your assignments.

BGP routing expertise is necessary. Your team needs to understand BGP operations, routing policies, prefix filtering, and basic network security. IXPs don’t provide technical support for configuring your routing – you’re expected to know how to peer properly.

Equipment capable of supporting your traffic volumes and BGP sessions is required. This typically means routers that can handle multiple 10+ Gbps ports and hundreds or thousands of BGP sessions. Underpowered equipment creates problems for you and potentially other IXP participants.

Physical Requirements

You need presence in the IXP’s facility. This means either colocating equipment in the facility where the IXP operates or establishing a connection from your facility to the IXP through a transport provider.

Most organizations choose colocation in IXP facilities because it’s simpler and typically cheaper than remote connections. Facilities in markets like Kansas City, Philadelphia, and Houston that host active IXPs provide straightforward access to these exchanges.

The equipment you install needs appropriate power and cooling, which is standard for data center colocation. A router with 2-4 ports drawing 500-1,000 watts is typical for IXP connectivity.

Policy Requirements

IXPs have acceptable use policies that participants must follow. Common requirements include:

No network layer filtering that would prevent traffic exchange with other participants. You can’t selectively block certain networks at the IXP level.

Maintain accurate routing information. Don’t advertise prefixes you don’t actually have the right to announce. Prefix hijacking and route leaks create problems for everyone.

Respond to operational issues in a timely manner. When other participants or IXP operations contact you about technical problems, you need to address them quickly.

Don’t use IXP connections for purposes beyond traffic exchange. The IXP port is for peering traffic, not for reaching your transit providers or establishing connectivity to other locations.

Membership Fees

IXPs charge various fees depending on their business model. Common fee structures include:

• Port fees based on the speed of your connection to the IXP.
• One-time setup fees, covering administrative costs and physical installation. 
• Some exchanges charge annual membership fees separate from port fees. This is more common in non-profit cooperative models where the membership fee represents your share of operational costs.
• Commercial IXPs sometimes offer different service tiers with premium features like enhanced support, remote peering options, or access to additional IXP locations under a single contract.

Cost Savings from IXP Connectivity

The financial benefits of IXP participation are substantial enough that connectivity pays for itself quickly for networks moving meaningful traffic volumes.

Transit Cost Reduction

The primary savings come from replacing expensive transit with near-zero marginal cost peering. Every gigabit of traffic you can route through IXP peering instead of transit eliminates ongoing per-megabit charges.

Calculate your potential savings:

1. Determine how much traffic you currently send through transit that could be peered
2. Multiply that by your transit pricing per megabit
3. Subtract IXP port costs

Even accounting for equipment costs, engineering time, and any facility expenses, the ROI is typically under six months for networks moving substantial traffic.

Bandwidth Efficiency

IXPs let you overprovision capacity without proportional cost increases. With transit, doubling your capacity doubles your costs. With IXP peering, upgrading from a 10 Gbps to 100 Gbps port might increase costs 2-3x while providing 10x the capacity.

This overprovisioning provides headroom for traffic growth without immediately triggering cost increases. Many organizations find their traffic grows into IXP port capacity over time without requiring frequent upgrades or cost increases.

Better Traffic Engineering

IXP connectivity provides alternative paths that improve overall network efficiency. When transit providers experience congestion or issues, IXP paths provide relief valves. This traffic engineering capability has value beyond direct cost savings by improving reliability and performance.

Some networks use IXP peering as their primary path for reaching many networks, keeping transit as backup. This “peering-first” strategy minimizes transit costs while maintaining transit for reaching networks you can’t peer with or for redundancy.

Real-World Examples

Content delivery networks report that IXP peering reduces their per-gigabit delivery costs by 60-80 percent compared to transit-only models. This cost reduction is what makes profitable video streaming possible at the massive scales users demand today.

Internet service providers use IXP peering to exchange customer traffic with content providers and other ISPs. Without IXPs, ISPs would need to pay transit costs to deliver content to customers, costs that would ultimately get passed to consumers through higher prices.

Enterprise networks with substantial internet traffic find IXP peering reduces costs while improving performance for cloud applications, SaaS services, and partner connectivity. Organizations that previously paid $100,000+ monthly for transit often reduce this to $20,000-30,000 through aggressive IXP peering.

Major IXPs and Regional Exchange Points

Internet exchange points operate globally with varying sizes and characteristics. Understanding the IXP landscape helps you identify relevant exchanges for your network.

Largest Global IXPs

DE-CIX Frankfurt operates the largest IXP by traffic volume, regularly exceeding 10 terabits per second. This reflects Frankfurt’s role as a major European internet hub and the exchange’s long history and strong participation.

AMS-IX Amsterdam is one of the oldest and most established exchanges, with hundreds of participants including major content providers, ISPs, and enterprise networks. Its neutral operation and central European location make it attractive to networks operating across the region.

LINX London serves as the UK’s primary exchange point with strong participation from European and international networks. London’s role as a financial and technology center drives demand for low-latency connectivity that LINX provides.

These large exchanges set standards for IXP operations that smaller regional exchanges follow. Their operational practices, policies, and technical approaches inform how IXPs operate globally.

Major U.S. Exchange Points

United States IXPs operate somewhat differently than European exchanges due to different market structures and greater distances between population centers.

Equinix operates exchange points in multiple U.S. markets including Silicon Valley, New York, Chicago, and others. These exchanges are commercially operated but maintain open participation policies.

NYIIX serves the New York metro area with strong participation from financial services networks, content providers, and regional ISPs. New York’s role as a major internet hub makes NYIIX relevant for networks serving East Coast users.

Midwest Internet Cooperative Exchange (MICE) in Minneapolis operates as a non-profit cooperative serving networks in the central U.S. It demonstrates how regional exchanges can serve areas underserved by large commercial IXPs.

Regional and Emerging Exchanges

Smaller regional exchanges serve specific geographic markets where local traffic exchange provides value. These might operate at much lower traffic volumes than major IXPs but still deliver benefits to local networks.

Kansas City IXPs provide regional exchange points serving the mid-country market. Networks at these exchanges exchange traffic locally rather than routing through coastal exchange points.

Philadelphia IXPs serve the mid-Atlantic region between New York and Washington. These exchanges benefit from Philadelphia’s role as an interconnection hub with strong carrier-neutral infrastructure.

Regional exchanges often start small but grow as more networks recognize the value. Early participants benefit most because they influence the exchange’s development and establish peering relationships before competition intensifies.

Selecting Relevant IXPs

When choosing which IXPs to participate in, consider:

Where your traffic originates and terminates. IXPs in markets where you have significant user or customer concentrations provide the most value.

Which networks you want to peer with. Different IXPs have different participant profiles. Research who participates at exchanges you’re considering to verify your target peers are present.

Geographic diversity for redundancy. Participating in IXPs in multiple markets provides alternative paths when issues affect specific locations.

Facility options and colocation costs. IXPs that operate in affordable facilities with good connectivity options provide better overall value than those in expensive or limited facilities.

Selecting Data Centers with IXP Access

Choosing colocation facilities with IXP access matters for networks planning to pursue peering strategies. Not all facilities host IXPs or make exchange participation straightforward.

IXP Presence in Facilities

Major IXPs operate in specific data centers where they’ve installed switching infrastructure. To participate, you need presence in one of those facilities or a connection from your facility to the IXP.

Some IXPs operate in multiple facilities within a market, providing flexibility in where you colocate. Others concentrate in single facilities, requiring you to colocate there if you want the lowest-latency, most cost-effective access.

When evaluating facilities, ask specifically about IXP access:

• Which IXPs operate in or connect to the facility?
• What’s involved in connecting to those IXPs?
• Are there additional costs beyond IXP port fees?
• How long does IXP connection typically take?

Remote Peering Considerations

Some IXPs offer remote peering where you don’t need physical presence at the IXP facility. Instead, you establish a connection from your facility to the IXP through a transport provider or remote peering service.

Remote peering trades off cost for convenience. You pay for the transport connection to reach the IXP, which typically costs more than local colocation near the IXP. But you avoid needing presence in multiple facilities if you want to participate in IXPs in different markets.

Latency increases slightly with remote peering compared to local presence. The transport connection adds distance and potentially hops. For most applications, this increase is minimal, but latency-sensitive applications might prefer local presence.

Carrier-Neutral Facilities

Facilities with carrier-neutral interconnection often host IXPs because the same characteristics that make facilities good for carrier-neutral environments make them attractive for IXP operations. Dense carrier presence, neutral operations, and established interconnection ecosystems support IXP development.

Markets with strong carrier-neutral facilities frequently develop active IXPs. The concentration of networks in these facilities creates the critical mass necessary for IXP success. Early IXP participants often choose these facilities specifically because they expect other networks to follow.

Multi-Market Strategies

Organizations serious about peering typically establish presence in multiple markets to participate in regional IXPs. This provides geographic diversity, reduces latency to users in different regions, and enables peering with networks that don’t participate in all exchanges.

A common strategy involves presence in 3-5 key markets covering major geographic regions. This might include a West Coast location, an East Coast location, a mid-country location, and perhaps international presence in Europe or Asia depending on your user base.

Strategic use of mid-country network hubs provides efficient aggregation points for peering. A facility in Kansas City with IXP access delivers balanced latency to both coasts while potentially offering lower costs than expensive coastal markets.

Interconnection Infrastructure Quality

Beyond IXP presence, evaluate the broader interconnection ecosystem at facilities. Strong interconnection capabilities enable not just IXP participation but also private peering and transit arrangements that complement your IXP strategy.

Look for facilities with:

• Multiple IXP options if available
• Numerous carriers for transit diversity
• Meet-me-rooms facilitating private peering
• Track record supporting network-heavy customers

Facilities specializing in network interconnection understand peering requirements better than general-purpose data centers. This expertise matters when you’re setting up complex peering architectures across multiple exchanges and private interconnections.

Ready to Reduce Network Costs Through Strategic Peering?

Internet exchange points represent one of the most effective ways to reduce network costs while improving performance. The savings from replacing transit with peering can run into hundreds of thousands or millions of dollars annually for networks moving substantial traffic. The performance benefits through lower latency and better routing provide additional value that directly improves user experience.

Organizations that don’t pursue IXP peering are leaving money on the table. The cost difference between transit and peering is so substantial that the investment in IXP participation pays for itself quickly. Getting started requires understanding IXP operations, securing appropriate facility access, and establishing peering relationships with relevant networks.

The network operators who benefit most from IXPs are those who recognize their value early and establish a presence before their traffic volumes make the cost savings obvious. By the time peering becomes clearly necessary, you’re behind competitors who’ve already optimized their costs through strategic exchange participation. Ready to explore IXP connectivity and peering strategies for your network? Netrality Data Centers operates facilities in markets with active internet exchange points and robust peering ecosystems. Our carrier-neutral interconnection infrastructure in Kansas City, Philadelphia, and Houston provides the connectivity foundation that serious peering strategies require. Contact our team to discuss your network requirements and explore how strategic facility selection in IXP markets can optimize your network costs and performance through effective peering relationships.