Network fundamentals

How Dedicated BGP Routing Protects Your Network

How routing policies, origin validation and alternative paths help protect connectivity, and what BGP cannot do by itself.

What BGP controls

Border Gateway Protocol, or BGP, lets independently operated networks exchange routes to IP address ranges. Those networks are identified by Autonomous System Numbers, or ASNs. Routing policy determines which announcements a network accepts and which paths it prefers.

Here, dedicated BGP routing means a routing design and policies agreed for your deployment. It is not a separate protocol or a guarantee of an exclusive Internet path. Clarify whether the service includes your own BGP session, provider-managed announcements or dedicated routing equipment.

Technical reference: IETF RFC 7454, BGP Operations and Security

Reduce the impact of routing mistakes

A routing error can advertise an address range from the wrong network or pass routes to a neighbour that should never receive them. Explicit import and export filters limit which prefixes are accepted and announced. Maximum-prefix limits help contain unexpectedly large route feeds.

BGP sessions also need restricted access and appropriate session protection. Changes should have a review and rollback process. These controls reduce exposure to mistakes and interference; their effectiveness depends on configuration and ongoing maintenance.

Operational guidance: RFC 7454, session protection and route filtering

Check who is authorised to announce a prefix

Resource Public Key Infrastructure (RPKI) lets an address holder publish a Route Origin Authorisation (ROA). It identifies an authorised origin ASN and permitted prefix lengths. Networks using Route Origin Validation can reject announcements that conflict with those records.

Origin validation does not verify the entire path or stop every hijack or route leak. A route with no covering ROA is not automatically malicious. Keep authorisations aligned with planned announcements so a legitimate routing change does not make your own routes invalid.

Further reading: RIPE NCC, BGP Origin Validation

Use alternative paths with realistic expectations

Connections to more than one upstream network can provide another route when one connection fails. Useful resilience also needs physical diversity: two services sharing a duct, router or power supply may fail together. Ask which failure scenarios the proposed design can survive.

A route change takes time to be detected and propagated. Existing sessions may be interrupted, and a reachable route does not prove that the application behind it is healthy. Test network failover alongside application monitoring and recovery.

You can control your outbound routing policy and influence inbound paths, but other networks make their own choices. BGP does not guarantee the lowest latency, a fixed geographical path or uninterrupted service.

Understand BGP’s role during a DDoS attack

With an agreed provider policy, a blackhole announcement can cause traffic to a targeted address range to be discarded upstream. This may relieve congestion affecting other services, but it also drops legitimate traffic to the target. It is a containment measure with an availability cost.

Traffic scrubbing and application protection require separate capabilities. Ask where filtering happens, how it is activated and which traffic is covered. BGP does not encrypt user traffic or replace firewalls, TLS, patching or access controls.

Technical reference: IETF RFC 7999, BLACKHOLE Community

What to agree with your network provider

Gloman operates AS50311 with infrastructure in Athens and Amsterdam. That provides a basis for discussing routing and connectivity directly with the team running the network. The exact filters, validation policy, mitigation facilities and response commitments for your service should be confirmed in its technical scope.

  • Who holds the address space, maintains its routing records and approves announcement changes?
  • Which failures are covered by alternative paths, and how will those paths be tested?
  • Who monitors changes in route visibility, and who responds to a suspected hijack?
  • Which actions can the NOC take during an attack, and what interruption might each action cause?

Why routing control and AS50311 matter to your business

An ASN identifies a network with its own routing policy. Operating AS50311 lets Gloman manage its announcements and upstream relationships directly. For a customer, the practical benefit is having an engineering team that can investigate and change the provider’s routing policy when a path causes problems.

Gloman peers at GR-IX. An Internet exchange allows participating networks to exchange traffic directly, potentially avoiding an unnecessary transit detour. The benefit depends on the destination network, available peering and the selected route. Membership does not mean every Greek connection follows the same path.

Compare the actual operator behind a hosting service, the escalation process and measured performance. Virtualisation does not determine routing quality: a reseller can use an excellent upstream, and owning an ASN alone does not guarantee a better path.

Further reading: GR-IX, the Greek Internet Exchange

Measure latency from the networks your users use

Test from representative Greek fixed and mobile networks, including regional offices and islands where relevant. Record round-trip time, packet loss, application response time and behaviour during a path failure, at normal and peak hours.

Treat a target such as below 10 ms as a requirement for named test locations and conditions, rather than a nationwide promise. Last-mile access, geography and remote routing policies all affect the result. Ask for measurements and a defined SLA before making a business commitment.