Our network: AS58269, own routing, redundant paths

AS58269 is the autonomous system of CloudHosting SIA, registered with RIPE NCC under the name AS-CLOUDHOSTING. An own ASN means we decide our routing policy ourselves: which carriers we buy transit from, how traffic enters and leaves, and how quickly we react when a path misbehaves. Providers without their own network rent all of this from someone else and can only forward your ticket when routing breaks. You can verify the whole setup in the public RIPE database: the AS, our prefixes and the route objects are all there.

Our own IPv4 allocations, including the 185.8.60.0/22 block and several /24 networks such as 91.220.43.0/24, all with route objects in the RIPE database and origin AS58269. On top of that we hold an IPv6 /32 allocation, 2a03:3440::/32. Addresses assigned to your services come from this space with correct whois data and reverse DNS under our control, which matters for mail deliverability and for anyone who checks where their provider's addresses actually come from. Whois for any of our addresses points at CloudHosting SIA in Riga, not at a chain of intermediaries, and service addresses are assigned from this same space without rented blocks.

Several independent carriers, deliberately unrelated to each other, each taking a full BGP feed; the exact set can change as we optimize routes, and we discuss details openly with colocation and enterprise clients. What matters for everyone else is the behaviour: if one upstream degrades or drops, BGP moves traffic to the remaining paths automatically within seconds, without an engineer having to wake up first. No single carrier failure becomes our outage, and that is the whole point of running an own AS.

Yes, in two ways. If you already hold PI addresses or an ASN, we announce your prefixes from our edge or set up BGP sessions with your AS; for colocation this is a standard add-on. If you do not have resources yet, our IP and AS registration service obtains PI IPv4, IPv6 and AS numbers for you via RIPE NCC, with CloudHosting acting as your sponsoring LIR and handling the paperwork. Your resources stay yours: if you ever leave, the addresses travel with you.

Three levels, by budget and requirement. A managed site-to-site VPN (IPsec or WireGuard) over existing internet connections is the quickest and cheapest. Carrier Ethernet gives you a point-to-point L2 circuit between your site and the data centre with no public internet on the path. Dark fiber across the Riga metro area is the top tier: your own strand, lit or unlit, with DWDM options when one wavelength is not enough. All three are designed, delivered and monitored by our engineers.

VPS plans come with a 1 Gbps port as standard and 2.5 or 10 Gbps options on bigger plans. Dedicated servers run on our 10 Gbit network. Colocation clients choose uplinks per rack, including 10G and guaranteed-bandwidth options from the add-on list. Inside the data centre, private VLANs between your own services are not billed as public traffic, so an app server talking to its database or backups flowing at night cost you nothing extra. If you are unsure what your workload needs, describe it to support and we will size the port honestly.

We hold our own IPv6 /32 allocation, 2a03:3440::/32, with a route6 object announced from AS58269; that is public RIPE data. IPv4 remains the default for services, because that is what most customer software still expects, and every service includes a dedicated IPv4 address. If your project needs IPv6, ask support when ordering and we will set it up for the service in question. We prefer saying this honestly instead of printing an IPv6 badge on every page of the site.

Because the address space is ours, reverse DNS is under our direct control: send the desired hostname to support and the PTR record is set for your IPv4 address. For mail servers this is essential, and we set PTR with your SPF, DKIM and DMARC story in mind, so deliverability checks pass consistently. Colocation and enterprise clients with their own prefixes keep control of their reverse zones, and we delegate rDNS to your name servers if you prefer to manage it yourself.

BGP does its job: sessions to the failed carrier drop, routes are withdrawn and traffic shifts to the remaining upstreams automatically. For most customers the event is invisible or a brief blip. Our engineers are alerted immediately and investigate the cause with the carrier while the network runs on the remaining paths. This is the practical difference between an own multi-homed AS and a single-uplink setup, where the same event means downtime until someone at the provider answers the phone.

Clearly malicious volumetric traffic is filtered at the network edge, and the network is monitored around the clock. For application-level attacks the server side matters most: a firewall, rate limiting and sane caching stop the bulk of what actually arrives, and our engineers help configure all of it. If your project is a likely target, talk to us before ordering: we will suggest an architecture with a protected frontend and backends on the private network, and involve an upstream scrubbing service where it genuinely makes sense.

From Riga, the Baltics are single-digit milliseconds away, the Nordic capitals and Warsaw sit within a few tens of milliseconds, and Western Europe is predictable because our routes ride major carrier backbones. We do not publish a synthetic latency table, because your real numbers depend on your users' networks, not on our marketing. The honest way to know: ask us for a test address to ping from where your users actually are, and measure before you buy. Sales that survive a ping test are the sales we like.

Yes, and you do not have to take our word for it. Open any RIPE database mirror and query AS58269: you will see AS-CLOUDHOSTING under CloudHosting SIA, our IPv4 route objects and the IPv6 /32 allocation. Run a traceroute to any of our services and it ends inside our own address space, not inside a reseller's network. This page states only what public records confirm, which is exactly how we would want any provider to talk to us. If anything on this page stops matching the records, tell us and we will fix the page, not the story.