Network AS47263: the BGP network underneath your clusters
In brief
AS47263 is Cube-Hosting's own autonomous system and the network that connects the managed clusters of managed-cluster.net. We operate Border Gateway Protocol (BGP) routing ourselves, buy transit from several independent carriers and peer at ERA-IX Frankfurt, NL-ix and frankonIX, among others. Your cluster gets static IP addresses, and you can check the routing yourself on bgp.tools.
AS47263 at a glance: key facts and where to verify them
You can verify the routing properties of AS47263 from the outside. The network's prefixes and peerings are publicly documented, for example on bgp.tools. For every fact, the table names the place where you can verify it or look it up.
Key facts AS47263 and where to verify each fact
Feature
Detail
Verify
Autonomous system
AS47263, operated by Cube-Hosting, RIPE region Germany
BGP network hosting: what AS47263 changes for your cluster
A hosting provider with its own autonomous system controls its routing. Through AS47263 we announce static IPv4 and IPv6 address space via BGP, over several independent upstreams and peering at internet exchanges. For you, that means static addresses with reverse DNS, routing you can check, and the option to bring your own PI addresses after review.
Static IP address space
Your cluster gets static IPv4 addresses from the AS47263 prefixes and, on request, dedicated IPv6 subnets, including reverse DNS. Static addresses can be entered reliably in your firewall rules, your partners' IP allowlists and your DNS records.
Redundant upstreams
AS47263 buys transit from several independent carriers. bgp.tools publicly documents which upstreams the network uses.
Peering in Frankfurt and beyond
At internet exchanges, AS47263 exchanges traffic directly with other networks, including at ERA-IX Frankfurt, NL-ix and frankonIX.
One network for clusters and servers
Managed servers run on the same AS47263 network as our clusters and come with IPv4 and an IPv6 /64. When a server grows into a cluster, we build the cluster in parallel and migrate with a rollback plan, on the same network.
Autonomous systems and BGP in brief
RFC 1930 defines an autonomous system (AS) as a connected group of IP prefixes run by one or more network operators with a single, clearly defined routing policy. Every AS has its own number; ours is 47263.
Autonomous systems use the Border Gateway Protocol (BGP) to tell each other which networks they can reach. RFC 4271 describes BGP as an inter-autonomous-system routing protocol; the primary function of a BGP speaker is to exchange network reachability information with other BGP systems.
managed-cluster.net is a brand of Cube-Hosting.net (owner: Jesse Röhsler, Gerolzhofen, Germany). We operate managed Proxmox and high-availability clusters in German data centers (Maincubes FRA01, NBG01 Nuremberg) on our own network, AS47263. How the brand, its owner and the network fit together is explained on our company page: Cube-Hosting.net, the operator of AS47263. The page on our FRA01 and NBG01 data centers explains the role of each site.
Sources
IETF: RFC 1930, Guidelines for creation, selection, and registration of an Autonomous System (AS), March 1996. rfc-editor.org↗, retrieved October 9, 2026.
IETF: RFC 4271, A Border Gateway Protocol 4 (BGP-4), January 2006. rfc-editor.org↗, retrieved October 9, 2026.
Cluster network
Cluster networking: redundant from core router to node
Inside the data center, the redundancy of AS47263 continues in the cluster: redundant switches and uplinks, bonding with LACP on every node, separate VLANs for management, storage and public traffic, a dedicated storage network and a redundant firewall and load balancer layer with health checks. We apply firewalling at every layer.
Diagram 01From upstream to node
Traffic from the internet
Link inside the data center
Storage network
external, not operated by us
active
Several upstreams and internet exchanges converge on the redundant core routers of AS47263. Below them sit the firewall and load balancer layer, the switches and the nodes, each of them redundant.Diagram as text
AS47263 connects to the internet through several independent transit carriers (shown as Transit A and B) and through peering at internet exchanges, including ERA-IX Frankfurt, NL-ix and frankonIX.
These connections terminate on redundant core routers (shown as Core 1 and Core 2). DDoS protection applies at the network level.
Behind them, a redundant firewall and load balancer layer with a floating IP and health checks distributes the traffic.
Redundant switches connect the nodes. Each node is attached through bonded uplinks (bonding with LACP).
Ceph replicates the nodes' data over a dedicated storage network.
Management, storage and public traffic run in separate networks (VLANs).
Redundant switches and uplinks: if a switch or a link fails, every node stays reachable over another network path.
Bonding with LACP on every node: a node's uplinks are combined into one logical connection.
Dedicated storage network: replication traffic for Ceph storage in the cluster runs on a network of its own, separate from user traffic.
Redundant load balancers: health checks take a failed node out of rotation, and a floating IP keeps your service address stable.
We define the network architecture, including VLAN separation, firewalls and redundant connectivity, during the design phase of your cluster and document it in writing. We discuss the hardware of the cluster nodes (CPU, RAM, storage, network) individually and define it in the design proposal. The Enterprise package adds separate networks, VPN, IP allowlisting and role-based access. How network, nodes and storage work together when something fails is explained on our high availability cluster page under redundancy at every layer.
DDoS protection
DDoS protection for hosting in Germany: at the network level, complemented inside the cluster
At managed-cluster.net, network-level DDoS protection is part of every fixed price, for managed clusters and managed servers alike. Inside the cluster, firewalling at every layer adds to it. If your application has special requirements for protection against flooding attacks, talk to us; we will tell you plainly what we can document today.
Protection layers From network to application: who secures which layer
Layer
Measure
Responsible
Network level
Network-level DDoS protection, included in every fixed price
Us
Cluster entry point
Redundant firewall and load balancer layer with health checks and a floating IP
Us
Networks inside the cluster
VLAN separation of management, storage and public traffic; firewalling at every layer
Us
Nodes and virtualization
Hardening with minimal services, restrictive firewalls, secured SSH access and network separation; security patches, critical vulnerabilities first
Us
Access
Separate networks, VPN, IP allowlisting and role-based access
Us (Enterprise)
Your application
You look after your application, we look after everything beneath it.
You
Our 24/7 monitoring covers the network as well; critical alerts reach on-call duty immediately. How network separation, patch management and backups support your security measures together is laid out on our security and compliance page: network security in the bigger picture.
IP addresses
IP addresses and IPv6 hosting: static addresses, reverse DNS, PI space
Every managed cluster gets static IPv4 addresses from the AS47263 prefixes and, on request, dedicated IPv6 subnets, including reverse DNS. IPv6 is native on our network, so no tunnels are involved. If you hold provider-independent address space (PI space), we can announce it via AS47263 after reviewing it. We size your address requirements during the design phase.
Addresses IP addresses at managed-cluster.net at a glance
announcement of your own provider-independent addresses via AS47263
after review, settled during the design phase
Term: PI space
Provider-independent (PI) addresses are assigned to your organization itself and are not tied to a provider. Even when we announce them via AS47263, they remain assigned to your organization.
FAQ
Network and AS47263: frequently asked questions
What does an autonomous system of your own mean for my cluster?
Routing we are responsible for ourselves, and static addresses. Through AS47263, we route your cluster's addresses over several independent transit carriers and peering at internet exchanges. Your cluster gets static IP addresses including reverse DNS, and you can follow the routing yourself on bgp.tools.
Can I bring my own IP addresses?
Yes, after review. If your organization holds provider-independent address space (PI space), we can announce it via AS47263. Whether and how that works for your addresses is settled during the design phase.
Do you support IPv6?
Yes. AS47263 routes native IPv6. Managed clusters get dedicated IPv6 subnets with reverse DNS on request, and managed servers come with an IPv6 /64 in addition to IPv4.
How does DDoS protection work?
Network-level DDoS protection is included in every fixed price. Inside the cluster, firewalling at every layer adds to it. If you have specific protection requirements, raise them with us before the project starts.
Where can I check the routing of AS47263?
On bgp.tools↗, which publicly documents the network's prefixes, upstreams and peerings.
Next step
What does your application need from the network?
Tell us about your address requirements, IPv6 needs, any PI space of your own and your rules for network separation. We take them into account when designing your cluster's network architecture and record the result in documentation that belongs to you.