1.01 epf-2026 2026-09-14 2026-09-16 3 00:05 https://2026.peering-forum.eu https://2026.peering-forum.eu/media/epf-2026/img/logo_2zDpVWM.webp Europe/Paris Plenary Talk 2026-09-14T09:30:00+02:00 09:30 00:20 Europe's digital markets are drifting toward dominant players. The usual answers are to merge or to compete. France-IX proposes a third: mesh. Concentration works by gravity: everyone stays because everyone stays. The question for anyone sitting in the middle of a network is whether they accumulate that gravity or spread it. We answer with three rules we impose on ourselves. We built the mesh in France first. Multipass gives a customer national reach from one port across six cities, and today roughly one bit in eight on an enabled port leaves for a city where that customer has no port. The same move now repeats between French exchanges, then across borders, through bilateral agreements. France-IX CEO Franck Simon shows what we have shipped, the rules we accept, and the first international partners, named on stage. epf-2026-103760-france-ix-don-t-consolidate-the-internet-mesh-it Franck Simon en false https://2026.peering-forum.eu/epf-2026/talk/LY9EZR/ https://2026.peering-forum.eu/epf-2026/talk/LY9EZR/feedback/ Plenary Talk 2026-09-14T09:50:00+02:00 09:50 00:20 Is the peering market slowing down, or are we simply using the wrong metrics to understand it? This talk argues that Internet peering is not in decline, but in transformation. Based on PeeringDB and ISOC Pulse IXP Tracker data, combined with long-term operational experience in the IXP ecosystem, it examines how Internet Exchange Points are evolving across different regions of the world. While some large and mature IXPs show slower growth in membership or public peering capacity, this reflects a broader shift in interconnection strategies rather than a weakening of the model. Traffic is becoming more concentrated, private interconnects and in-network caching are increasingly complementing public peering, and edge architectures are changing where networks exchange traffic. The talk will highlight the factors that determine IXP growth and sustainability, including telecom market liberalisation, local content demand, data centre ecosystems, and national infrastructure policy. It will show why IXPs remain strategic assets for digital resilience, local traffic exchange, and a more robust, decentralised Internet. epf-2026-100685-peering-market-at-a-glance-trends-transformations-and-regional-dynamics-of-internet-interconnection Flavio Luciani en false Paper Slide presentation (draft) Final slides PPTX Final slides PDF https://2026.peering-forum.eu/epf-2026/talk/XNCE89/ https://2026.peering-forum.eu/epf-2026/talk/XNCE89/feedback/ Plenary Talk 2026-09-14T10:10:00+02:00 10:10 00:20 The distribution of CDN, cloud and content provider capacity across Internet Exchanges (IXPs) provides a fascinating lens into the physical infrastructure of the Internet and public peering. While hyperscalers and CDNs increasingly deploy private backbone infrastructure, public peering remains a critical mechanism for regional traffic distribution, cache efficiency, eyeball reach, and resilience. This study aggregates the port capacity of ten of the most IXP-interconnected CDN, cloud and content networks and ranks IXPs by total capacity. This ranking provides insights into the choices made by the networks, on an individual basis but also collectively as a group of CDN, cloud and content providers with significant influence in the global interconnection ecosystem. The networks covered are Akamai, Meta, Amazon, Hurricane Electric, Cloudflare, Fastly, Microsoft, Google, Netskope and ACE CDN. The limitations of this analysis are obvious: presence at exchanges could be incorrect or port capacity could be out of date, for instance. However, most of the networks under analysis depend heavily on PeeringDB automation. This gives the report a relatively high level of confidence. epf-2026-103130-peering-capacity-at-public-internet-exchanges-what-the-data-reveals Gael Hernandez en false Study published in RIPE Labs Slides final https://2026.peering-forum.eu/epf-2026/talk/DYTACZ/ https://2026.peering-forum.eu/epf-2026/talk/DYTACZ/feedback/ Plenary Talk 2026-09-14T11:00:00+02:00 11:00 00:20 The Internet is not moving away from interconnection, but the way networks interconnect is changing. For decades, Internet Exchange Points have served as visible, open, and locally grounded meeting points where networks could exchange traffic, build relationships, improve performance, and strengthen regional Internet ecosystems. Today, that model remains essential, but it is increasingly accompanied, and sometimes bypassed, by other forms of interconnection: Private Network Interconnects, remote peering, Virtual Peering Providers, cloud on-ramps, CDN deployments, and hyperscaler interconnection fabrics. This presentation explores the “Peering Disconnect”: the growing gap between the traditional understanding of peering as open, public, local, and many-to-many, and the evolving reality of traffic exchange that is increasingly private, concentrated, platform-mediated, and less visible to the wider community. The session will examine how these shifts affect IXPs, network operators, content networks, and smaller providers seeking efficient and equitable access to major traffic sources, while taking into consideration differences between connectivity developed and developing markets. Rather than asking whether IXPs are still relevant, the presentation asks how their role is changing. Are IXPs still primarily traffic aggregation points, or are they becoming trust anchors, discovery platforms, community institutions, resilience infrastructure, and neutral convening spaces in a more complex interconnection environment? What happens when the most important traffic relationships move into private arrangements? And how should the peering community explain these changes to policymakers who may already misunderstand how Internet interconnection works? Drawing on ideas from recent Internet infrastructure policy work, the presentation will also consider why these market and technical shifts matter for regulatory debates. Attendees will leave with a sharper understanding of how peering is evolving, what risks are emerging, and what the interconnection community can do to preserve openness, resilience, and local value. epf-2026-103714-the-peering-disconnect-rethinking-ixps-in-the-age-of-pnis-and-cloud-interconnection Hanna Kreitem en This session explores the “Peering Disconnect”: the widening gap between the traditional understanding of peering as open, public, local, and many-to-many, and the evolving reality of traffic exchange through PNIs, VPPs, cloud platforms, remote peering, and hyperscaler interconnection fabrics. It will examine what these shifts mean for IXPs, network operators, smaller networks, and policymakers, and how the peering community can preserve openness, resilience, and local value in a more private and platform-mediated interconnection environment. false Final Slides https://2026.peering-forum.eu/epf-2026/talk/ZGEELL/ https://2026.peering-forum.eu/epf-2026/talk/ZGEELL/feedback/ Plenary Talk 2026-09-14T11:20:00+02:00 11:20 00:20 Modern DDoS attacks are becoming more distributed, encrypted, and difficult to mitigate. While large networks often deploy dedicated protection platforms, many smaller networks connected to IXPs lack effective DDoS defenses. At the same time, Internet Exchange Providers occupy a unique position within the Internet ecosystem, offering unparalleled visibility into traffic flows and attack patterns. This presentation explores how IXPs can evolve beyond traditional peering services by offering security capabilities to their members. We will review emerging attack trends, discuss the network architectures required to support scalable DDoS detection and mitigation services, and examine how DDoS protection can both improve member security and create new value-added service opportunities for IXPs. Attendees will gain practical insights into attack evolution, architectural design considerations, and the role IXPs can play in building a more secure and resilient Internet. epf-2026-103271-the-attack-is-coming-through-your-peering-ports /media/epf-2026/submissions/RVKH77/image_llvQ17H.webp Dominique Schaeffer en false The attack is coming through your peering ports ! https://2026.peering-forum.eu/epf-2026/talk/RVKH77/ https://2026.peering-forum.eu/epf-2026/talk/RVKH77/feedback/ Plenary Talk 2026-09-14T11:40:00+02:00 11:40 00:20 PeeringDB is shaped not only by its users, but by those who take part in its governance. This session explains how PeeringDB’s committees work, how members of the interconnection ecosystem can contribute their expertise, and how the pdb-gov mailing list provides a direct channel into discussions on the organization’s priorities, evolution, and future direction. epf-2026-103536-peeringdb-what-s-next /media/epf-2026/submissions/TBT9T3/image_ZVDWV6K.webp Livio Morina en • PeeringDB is more than a database. It is a community-driven, member-based non-profit organization. • Governance is not managed only by the Board. Much of the work happens through committees and community participation. • Governance structure: Members. Board of Directors. Committees. Staff. Volunteers. • The role of the committees: Admin Committee. Operations Committee. Product Committee. Outreach Committee. • What each committee is responsible for, and how each contributes to PeeringDB's evolution. • Why committee participation matters: technical expertise from operators, industry perspective, community consensus, and transparent decision-making. • How ideas become reality: community feedback. GitHub issues. Committee discussion. Board oversight where appropriate. Implementation. • Ways to contribute without joining the Board: join a committee, participate in discussions, review proposals, help define priorities, contribute documentation, support outreach activities. • PeeringDB Membership: who can become a member, member rights and responsibilities, Board elections, Member Meetings. • The pdb-gov mailing list: the primary forum for governance discussions, bylaws, elections, strategic direction, committee activities, governance proposals, and major organizational decisions. • Why subscribe to pdb-gov: stay informed before decisions are made, participate in discussions, influence the future direction of PeeringDB. • Transparency, openness, and community consensus are fundamental principles of PeeringDB governance. • Call to action: become a Member, subscribe to pdb-gov, join a Committee, and help shape the future of PeeringDB instead of simply using it false FINAL https://2026.peering-forum.eu/epf-2026/talk/TBT9T3/ https://2026.peering-forum.eu/epf-2026/talk/TBT9T3/feedback/ Plenary Talk 2026-09-15T09:00:00+02:00 09:00 00:20 The Digital Networks Act is being legislated now, and it proposes to bring in new requirements for IP interconnection relationships (and much else besides). This talk provides an update on the text as it stands, what it might mean for peering and content delivery relationships, and how you can provide feedback to policymakers as the proposals develop. epf-2026-103525-the-digital-networks-act-implications-for-ip-interconnection Mike Blanche en false https://2026.peering-forum.eu/epf-2026/talk/ZDBJ3F/ https://2026.peering-forum.eu/epf-2026/talk/ZDBJ3F/feedback/ Plenary Talk 2026-09-15T09:20:00+02:00 09:20 00:20 The European Commission’s proposal for a Digital Networks Act (DNA) largely preserves the current approach to IP interconnection in the EU, despite the intense political debate on “fair share” and sender‑pays models that preceded it. Interconnection rules remain confined to public electronic communications networks and services, in continuity with the EECC and Open Internet Regulation, and the DNA opts for voluntary ecosystem cooperation and conciliation mechanisms on IP interconnection rather than imposing network usage fees or mandatory settlement regimes. This outcome is broadly positive for the openness and contestability of European interconnection markets, but the regulatory trajectory is not yet settled, as the proposal will be subject to institutional negotiations and possible re‑balancing of interests. At the same time, the draft DNA introduces a broader jurisdictional hook over “digital infrastructure” in general – notably through provisions such as Article 9(2) – which may extend telecom‑style oversight to infrastructures and private networks that have so far remained largely outside sector‑specific regulation. One plausible consequence is an expansion of regulatory leverage over private interconnection arrangements, including CDNs and data centre‑based ecosystems, without a clear articulation of the policy objective, especially in light of overlapping obligations under NIS2 for network and information‑system security. Against this background, the presentation will critically assess the evolution of the EU interconnection framework, contrasting the institutional choice against “fair share” with emerging forms of soft‑law and voluntary cooperation, and exploring the risks of an incremental extension of jurisdiction over private networks. It will do so in dialogue with recent empirical work on the changing peering and interconnection landscape – including evidence on the growth of IXPs, private interconnects and CDNs – to test whether the DNA’s calibrated intervention is adequate to address current and future market developments. epf-2026-102953-evolution-of-the-european-regulatory-framework-on-interconnection-and-peering-prospects-and-risks Innocenzo Genna en The proposed presentation examines the European Commission’s Digital Networks Act (DNA) proposal from the perspective of IP interconnection and the broader evolution of the EU regulatory framework. Its core argument is that the current DNA proposal does not alter the existing European approach to interconnection in any fundamental way, since regulation remains centred on public electronic communications networks, broadly in continuity with the current framework. This is an important and constructive outcome in light of the earlier policy debate around “fair share” and sender-pays mechanisms, which the proposal does not introduce. At the same time, the presentation aims to highlight areas that deserve close scrutiny during the legislative process, including the possible implications of the DNA’s broader jurisdictional reach over digital infrastructure and the uncertain boundaries between telecom regulation, private networks, and existing security obligations such as NIS2. The analysis is also informed by recent market evidence on peering, IXPs, and the increasing relevance of private interconnection arrangements in Europe and beyond. The topic is intended for an audience of peering, policy, and regulatory professionals, and is designed to contribute a legal and policy perspective to current market and technical discussions at EPF. It is hoped that the presentation will support a clearer understanding of both the opportunities and the risks associated with the next phase of European connectivity regulation. The presentation will consist of a single talk of 20-30 minutes, with slides. false Final slides https://2026.peering-forum.eu/epf-2026/talk/YVS8DH/ https://2026.peering-forum.eu/epf-2026/talk/YVS8DH/feedback/ Plenary Talk 2026-09-15T09:40:00+02:00 09:40 00:20 In 2014 Liberty Global PLC's acquired Ziggo NV in the Netherlands, merging it with UPC. The European Commission gave a conditional clearance decision. One of the remedies offered to address market concerns was a commitment to maintain sufficient direct interconnect capacity until May 2026. epf-2026-103533-case-m-7000-wrap-up-of-a-commitment-in-the-ipx-domain Steven Van Steen en Demystifying some aspect of how major M&A get approved in Europe. Overview of the remedy in the IPX (peering) domain: how it was organized and addressed which concerns, what unexpected challenges and side effects given evolving reality over twelve years. What could have been done better. false EU Competition Case M.7000 Link to last public report Case M.7000 wrap-up of a commitment in the IPX domain (PDF) https://2026.peering-forum.eu/epf-2026/talk/S8ENBA/ https://2026.peering-forum.eu/epf-2026/talk/S8ENBA/feedback/ Plenary Talk 2026-09-15T10:00:00+02:00 10:00 00:20 The content delivery landscape has shifted substantially since our 2023 presentation titled "Embedded CDNs in 2023﹣A status", and the pace of change shows no sign of slowing. This talk revisits the CDN ecosystem three years on, focusing on three intertwined developments: structural change in static content delivery, the rise of live streaming as a dominant traffic class, and the resulting operational pressure on service providers. On the static side, the market has consolidated and reorganized. Dominant legacy players are now delivering a significantly larger share of their traffic through a combination of on-net caches and their own backbone networks, while several previously prominent CDNs have either exited the market or been absorbed through acquisition. Open-caching models that had begun to mature now face an unclear future amid this shifting landscape. Meanwhile, a new wave of entrants is reshaping how edge capacity is sourced, deployed, and monetized — challenging incumbents and introducing novel approaches to edge infrastructure In the live streaming segment we see continued growth in the number of live events that are streamed to ever increasing audience sizes. In this talk we'll examine what operators are doing to keep up and if their methods are effective. And finally, it wouldn't be 2026 without a look into the growth of AI services and we'll look to see if the massive growth rates has any impact on content delivery. epf-2026-103777-the-cdn-landscape-revisited-static-live-and-ai Jac Kloots en false https://2026.peering-forum.eu/epf-2026/talk/GRDT3B/ https://2026.peering-forum.eu/epf-2026/talk/GRDT3B/feedback/ Plenary Talk 2026-09-15T10:20:00+02:00 10:20 00:20 The AI infrastructure market is at a critical inflection point. The "Neocloud" sector is experiencing massive growth, forecasted to hit $20 billion in revenue by 2026. However, these independent GPU operators are caught in a "commodity trap". Competing purely on raw compute pricing against dominant hyperscalers—who hold over 60% of the market—is an unsustainable model, especially given that AI hardware depreciates 3 to 5 times faster than traditional equipment. To survive, infrastructure providers must evolve from basic bare-metal providers into Sovereign Cloud operators. This evolution is fundamentally driven by the market shifting from AI Training to AI Inference. By 2030, the AI inference market is projected to reach $150 billion, overtaking training revenues. While AI training relies on bulk data transport and high bandwidth, inference operates as a real-time "nervous system" demanding strict latency, resilience, and predictability. The core challenge our industry must address is that best-effort IP transit is fundamentally unfit for enterprise-grade sovereign AI inference. Relying on the public internet offers zero control over data paths, no observability into exchange points, and highly unpredictable latency. This presentation argues that the technical enabler for the next generation of AI infrastructure requires abandoning standard IP Transit in favor of strategic Peering at Neutral Internet Exchanges (IXs). Key Discussion Points:Regaining Network Control: How shifting to neutral IXs provides fine-grained control over routing, reduces hops, improves latency, and limits reliance on single network providers. Sovereignty as a Competitive Moat: With proposed legislation like the EU Cloud and AI Development Act (CADA), data sovereignty is moving from an enterprise demand to a strict legal requirement. Peering enables "jurisdiction-aware interconnection," guaranteeing data residency in ways that opaque hyperscaler routing simply cannot match. The AI-Native Stack: Transitioning from a pure hardware business to a unified stack that combines Bare Metal, Sovereign Connectivity & Peering, and AI-Native Software. Ultimately, connectivity is no longer just the plumbing; it is the defining characteristic of the next-generation Neocloud. This session will outline why a mature peering strategy is the only way to escape the hardware commodity trap and capture the future of AI inference. epf-2026-102175-neocloud-evolution-escaping-the-commodity-trap Kaj Kjellgren en false Latest version of the presentation https://2026.peering-forum.eu/epf-2026/talk/JHQAVK/ https://2026.peering-forum.eu/epf-2026/talk/JHQAVK/feedback/ Plenary Talk 2026-09-15T11:10:00+02:00 11:10 00:20 LON2 was originally launched in 2002 following LON1’s increasing popularity and was created to ensure there was no critical single point of failure in the UK’s internet connectivity. LINX’s dual LAN infrastructure design in London has been something that sets us apart over the years from other Internet Exchange Points. Operating a redundant, diverse and resilient solution in house for members means not having to look elsewhere for this critical capability. In 2018 LINX announced that they were the world’s first IXP to deploy a disaggregated network using EVPN, with hardware and software sourced from alternative vendors. However, with end of life equipment, how did LINX refresh a LAN with 1/3 of its membership on without issues... or was there? epf-2026-103463-lon2-network-refresh-linx-metro-resilience Mariano Juliá en false https://2026.peering-forum.eu/epf-2026/talk/K7TB3M/ https://2026.peering-forum.eu/epf-2026/talk/K7TB3M/feedback/ Plenary Talk 2026-09-15T11:30:00+02:00 11:30 00:20 The Traffic Radar working group focuses on establishing an open and partially anonymous mechanism for information exchange between content providers, CDNs, and internet service providers regarding expected peaks in internet traffic. Its primary objective is to enable organizations to anticipate and prepare for significant traffic surges, thereby reducing network strain and service disruptions. epf-2026-101005-traffic-radar-tsunami-warning Rob de RuigJan van BoesschotenSteven Van Steen en The initiative is driven by increasingly complex and frequent traffic spikes, caused by overlapping large-scale digital events such as game updates, live streaming, and emerging AI-generated traffic. These dynamics are expected to intensify in the coming years, further increasing pressure on internet infrastructure. false Final presentation - pdf https://2026.peering-forum.eu/epf-2026/talk/BR7UST/ https://2026.peering-forum.eu/epf-2026/talk/BR7UST/feedback/ Plenary Talk 2026-09-15T11:50:00+02:00 11:50 00:20 At Internet Exchange Points, equal-cost multipath is the norm — but equal-cost is rarely equal capacity. An IXP member with a 10G and a 100G port to the same route server, or a network with multiple upstream exit links of 800G, 400G, and 200G, gets traffic distributed evenly across all paths regardless of what each can actually carry. The result is chronically oversubscribed low-capacity paths and underutilized high-capacity ones. This problem is well understood operationally, but until recently lacked a standardized and interoperable BGP-native mechanism to express it. RFC 10005, "BGP Link Bandwidth Extended Community," published as an Internet Standard in June 2026, addresses this directly: a BGP extended community encoding link bandwidth in bytes per second that receiving speakers use to compute proportional, weighted load-balancing across ECMP next hops. Its applicability is broad — from asymmetric IXP port configurations and multi-homed peering across exchanges of different capacity, to content providers with WAN uplinks of unequal speeds. This talk examines RFC 10005 alongside its companion use-cases document (draft-ietf-bess-ebgp-dmz) from both the protocol and operator perspective. We will cover the encoding format, the transitive versus non-transitive community variants and their significance at IXP route servers and peering boundaries, and re-advertisement procedures when next hops change. We will walk through production implementation features — including auto-sensed link bandwidth and aggregate-bandwidth computation across ECMP paths — showing how these reduce manual configuration overhead at scale. The community has also been implemented in production routing platforms, providing concrete validation of deployment readiness. We will also address the peering policy question of when to strip the community to avoid leaking internal capacity information beyond the intended administrative domain. The talk will also aim to include a customer perspective, with real-world deployment experiences from operators being sought for the final presentation. epf-2026-103726-rfc-10005-bgp-link-bandwidth-extended-community-standardization-of-bgp-weighted-ecmp Reshma Das en Learning Objectives 1. A concrete understanding of how RFC 10005 solves the asymmetric capacity problem at IXP peering edges — and why plain ECMP falls short when members have ports or exit links of different speeds. 2. Practical guidance on the transitive versus non-transitive community variants, including their behaviour at IXP route servers and whether operators should filter the community before it reaches bilateral peers or transit providers. 3. An understanding of how multi-homed IXP members and content providers can achieve proportional traffic distribution across asymmetric peering capacity without requiring changes to the route server itself. 4. Familiarity with production implementation features — auto-sensed link bandwidth and aggregate-bandwidth computation —and how they map to the RFC mechanisms, reducing manual operational overhead at scale. 5. Real-world deployment outcomes where available, and a concrete reference showing the community producing measurable, proportional forwarding behaviour. false BGP LBW RFC BGP LBW PPT https://2026.peering-forum.eu/epf-2026/talk/7ATQHH/ https://2026.peering-forum.eu/epf-2026/talk/7ATQHH/feedback/ Plenary Talk 2026-09-16T09:40:00+02:00 09:40 00:20 Turkiye is a geographically large and populous country bridging Europe, Asia and the Middle East. Turkiye has high rates of digital literacy, mobile and fixed broadband usage, and draws considerable traffic flows. Many neighboring countries to the east and south connect to Turkey for internet connectivity. Despite this, the IXP ecosystem remains relatively underdeveloped, with significant traffic being exchanged in Sofia rather than within Turkiye. We will present an overview of the Turkish internet ecosystem: • Operators and market structure • Infrastructure realities (access, metro, and intercity networks) • Data centers and Hosters • Content providers and Cloud players • IXPs, cybersecurity solution providers, etc • Licensing regime, regulatory and relevant government authorities • Domestic and international Traffic flows and discuss needs and opportunities as well as possible ways forward. The regulator has placed IXPs on its work program and is actively soliciting views on how to help foster domestic IXPs. The regulator is especially keen to retain certain sensitive and / or essential traffic flows within the country’s borders to ensure digital sovereignty and digital continuity. Given the geography and size of Turkiye as well as traffic pull from eastern and southern neighbors, we will investigate regional peering models, chicken-egg issues regarding data centers outside of the main 2-3 cities, and potential involvement of the NREN Ulakbim (Geant member) for colo and administration. We will stress the potential wider value for the domestic ecosystem of thriving IXPs, such as its positive impact for fostering security, content and other value-added start-ups and providers. Following the presentation, both in Q&A as well as in coffee breaks, we will solicit and seek to benefit from the views, advice and wisdom of participants. epf-2026-103427-the-internet-ecosystem-in-turkiye-needs-opportunities-and-ways-forward-for-ixp-s Cem Çelebiler en false Final slides https://2026.peering-forum.eu/epf-2026/talk/BZXRAE/ https://2026.peering-forum.eu/epf-2026/talk/BZXRAE/feedback/ Plenary Talk 2026-09-16T10:00:00+02:00 10:00 00:20 Route-server looking glasses already provide valuable visibility into routing at Internet Exchange Points, either by querying the routing software directly, as Alice Looking Glass does, or through dedicated BGP collection sessions. They help operators and participants inspect exchanged prefixes, troubleshoot connectivity, and better understand route-server activity. In this presentation, we introduce BMP as an alternative and complementary approach to monitoring IXP routing. BMP can provide visibility into pre-policy, filtered, and non-best routes, together with detailed information about peers and session state. Both BIRD and FRRouting can export BMP data when deployed as route servers, although their current implementations have limitations and do not support every BMP monitoring view. We then present the new bgproutes.io IXP Explorer. By combining BMP data from route servers with routing data contributed by IXP participants, it provides real-time and historical visibility into routes exchanged through both route servers and bilateral sessions. The platform also includes RPKI ROA and ASPA validation results. All data is openly accessible through the IXP Explorer dashboard and the bgproutes.io public API for programmatic analysis. epf-2026-103659-expanding-routing-visibility-at-ixps-with-bmp Thomas Holterbach en Visibility into the routes exchanged at an Internet Exchange Point is critical. It helps IXP operators monitor routing activity and enables participating networks to understand which prefixes are reachable through the exchange, whether through route servers, bilateral peering sessions, or both. In this presentation, we will first examine what IXP operators and participants require from a BGP route-monitoring system. Networks need to determine which prefixes are exchanged, which BGP attributes are attached to them—such as AS paths and communities—and through which sessions they are visible. Access to historical data is also valuable for investigating routing events, assessing participant behaviour, and identifying potential configuration problems. RPKI ROA and ASPA validation results can provide additional insight into the security of the exchanged routes. We will then review the main approaches currently used to provide public visibility into IXP routing. These include collecting routes through a dedicated BGP session with a route server and querying the route-server software directly, as implemented by tools such as Alice Looking Glass. We will discuss the strengths of these approaches, as well as the visibility that remains unavailable when compared with the monitoring requirements identified above. Next, we will introduce the BGP Monitoring Protocol, or BMP, and explain why it is well suited to monitoring routing at an IXP. BMP is a standardised protocol that provides a continuous stream of routing information and peer-state changes. By exposing routes before and after inbound policy processing, it can reveal routes received from each participant, routes accepted by the route server, routes removed by inbound filtering, and accepted routes that were not selected as the best path. It also provides metadata about monitored BGP sessions, including peer addresses and session-up and session-down notifications. We will also discuss the limitations of current BMP implementations in widely used route-server software. BIRD’s and FRRouting provides Adj-RIB-In pre-policy and post-policy, but does not currently document support for BMP Adj-RIB-Out. Finally, we will introduce the new IXP Explorer developed by bgproutes.io. The platform allows operators and participants to: * View routes received by route servers connected to bgproutes.io through BMP, including filtered routes and accepted routes that were not selected as the best path. * View routes exchanged over bilateral sessions when IXP participants also connect their routers to bgproutes.io. * Explore current and historical RIB snapshots and BGP updates, together with RPKI ROA and ASPA validation results. * Access the underlying data programmatically through the bgproutes.io public API. false Final slide deck https://2026.peering-forum.eu/epf-2026/talk/LTLSJQ/ https://2026.peering-forum.eu/epf-2026/talk/LTLSJQ/feedback/ Plenary Talk 2026-09-16T10:40:00+02:00 10:40 00:20 Most peering and transport networks are moving from 400G ZR+ to ULH reaches and on to 800G. The whole wavelength is ours — so it's worth asking the optical layer a question we rarely do: is it safe? The bulk cipher we rely on, AES-256, stays strong even against a quantum computer; the exposed element is the key exchange. And "harvest now, decrypt later" — recording encrypted traffic today to decrypt it once a quantum computer exists — makes that a present-day problem for any data with a long secrecy life. This talk shows a constructive, non-disruptive fix: quantum key distribution feeds quantum-safe keys into the MACsec or IPsec encryptors you already run, via the standard ETSI GS QKD 014 interface — a boost, not a forklift. We close with the physics of running a quantum channel on live DWDM: Raman coexistence, guard bands, and trusted-node distance limits. epf-2026-102011-from-400g-zr-to-ulh-800g-strengthening-network-encryption-with-qkd Marcin Bała en false Final slides in PDF Final slides in PPTX https://2026.peering-forum.eu/epf-2026/talk/PZ7DXA/ https://2026.peering-forum.eu/epf-2026/talk/PZ7DXA/feedback/ Plenary Talk 2026-09-16T11:00:00+02:00 11:00 00:20 In the old days of traditional separation of IP and DWDM, there was no need for IP folks to care about much other than what they used on their routers/switches (while the transport team ran the DWDM infra). These past DWDM infra approaches for maximizing the data capacity per fiber pair went for running more and more DWDM channels with grid spacings as small as possible. This meant that grid spacings shrank from 200GHz to 100GHz and then 50GHz with some applications even going for 25GHz. Now in the age of IPoDWDM the bandwidth per channel keeps increasing, as complex modulation schemes come into favor over ON-OFF-Keying which has been the de facto standard so far. Those increased per channel bandwidths of 400Gbps, 800Gbps and now pushing into the 1.6Tbps realm demand for larger grids to accommodate the spectrum necessary to operate such “Superchannels”. Especially the fact of coherent detection being “blind” to anything but its own wavelength has enabled interesting topologies that can omit filters altogether. Of course that comes at a cost of reduced flexibility. Which means, IP folks now will need to understand the nuances of how coherent detection works, need for larger filters for 800Gbps and higher, and how 400GHz filters (key word here being COTS) would future proof your network by not only allowing you to grow to 1.6Tbps, but also allow backwards compatibility for legacy services on 50GHz, etc. Made by FLEXOPTIX Research - Gert and Thomas. epf-2026-96205-channelmania-future-proof-your-dwdm-network-topology-while-keeping-it-flexible-for-1-6t /media/epf-2026/submissions/ENNGJW/image_2dGJNd2.webp Thomas Weible en false FLEXOPTIX Channelmania + INC bonus https://2026.peering-forum.eu/epf-2026/talk/ENNGJW/ https://2026.peering-forum.eu/epf-2026/talk/ENNGJW/feedback/ Plenary Talk 2026-09-16T11:20:00+02:00 11:20 00:20 IP over DWDM (IPoDWDM) is not a new idea. Its first major wave of adoption emerged around 2009, promising to simplify network architectures by integrating optical transport directly into IP routers. However, the technology struggled to gain traction. Large, expensive line cards, high power consumption, fixed optical interfaces, proprietary vendor implementations, and operational boundaries between IP and optical teams ultimately limited its commercial and operational appeal. More than fifteen years later, the conditions have fundamentally changed. The emergence of high-performance coherent pluggable optics has transformed the economics and practicality of IPoDWDM. Compact, interoperable 400G and emerging 100G and 800G coherent modules now allow routers to connect directly across DWDM infrastructure without traditional transponder layers, reducing equipment, power, space, and operational complexity. This presentation explores why IPoDWDM failed the first time, why the technology is increasingly viable in 2026, and what has changed across the optical ecosystem. It also examines how open line systems, optical disaggregation, standardized interfaces, and software-driven control frameworks are enabling operators to rethink the relationship between IP and optical networks. Rather than simply revisiting an old architecture, modern IPoDWDM creates opportunities for simpler, more flexible and economically efficient network designs. epf-2026-96531-ipodwdm-in-2026 Mark Tinka en false IPoDWDM in 2026 https://2026.peering-forum.eu/epf-2026/talk/JAYNEN/ https://2026.peering-forum.eu/epf-2026/talk/JAYNEN/feedback/ Plenary Talk 2026-09-16T11:40:00+02:00 11:40 00:20 This presentation addresses a critical strategic vulnerability in telecommunications infrastructure: the hidden costs and compounding failures of prioritizing low upfront costs when deploying Internet Service Provider (ISP) backbone networks. While budget-grade optical transceivers, cut-rate leased fiber paths, and under-provisioned routing hardware appear to optimize initial Capital Expenditure (CapEx), they systematically degrade network integrity. Technically, this manifest as small buffer bottlenecks, TCAM memory exhaustion under global BGP loads, and high packet drop rates during peak traffic intervals. Beyond immediate performance deficits, the presentation explores two critical, non-technical vectors of degradation: human capital and external peering ecosystems. Internally, a substandard network fosters a reactive "firefighting" culture, accelerating engineering burnout and triggering a costly talent drain of top network architects. Externally, underpowered infrastructure destabilizes peering sessions at public Internet Exchange Points (IXPs), leading to route flapping, port quarantine by IXP operators, and expensive reliance on fallback IP Transit links. Ultimately, the short-term savings of cheap infrastructure are rapidly erased by severe Operational Expenditure (OpEx) inflation, steep Service Level Agreement (SLA) penalties, regulatory fines, and permanent customer churn. The presentation concludes with a five-year strategic investment framework, advocating for geometrically diverse fiber redundancy, automated MPLS/SRv6 failover protocols, and robust edge-router sizing to protect the network's Total Cost of Ownership (TCO) and ensure long-term operational resilience. epf-2026-103305-cheaper-doesn-t-mean-better Marco Paesani en false Cheaper doesn't mean better (Final) https://2026.peering-forum.eu/epf-2026/talk/SQZZTZ/ https://2026.peering-forum.eu/epf-2026/talk/SQZZTZ/feedback/