South Africa's premier carrier-neutral network with triple-redundant IP transit and extensive peering across major internet exchanges
Rackzar operates AS329206, providing robust, redundant connectivity across South Africa's premier data centres
Rackzar's autonomous system (AS329206) is a Network Service Provider operating across South Africa with an open peering policy. Our network is designed for reliability, performance, and scalability.
We maintain an open peering policy to ensure optimal routing and performance for our customers. Our presence at major South African internet exchanges enables direct, low-latency connections.
Engineered for uptime: layered network security, RPKI-validated routing, strict prefix filters and diverse upstreams. Private VLAN isolation and a dedicated management plane with role-based access.
Rackzar takes IP transit from three independent upstream providers: Workonline, Agile and Network Platforms. Each carrier reaches us over separate physical paths and separate BGP sessions, so a fibre cut, a maintenance window or an upstream incident on any one of them does not interrupt your traffic.
Most South African hosts run one or two upstreams. Running three means we can lose a carrier entirely and still hold full redundancy on the remaining two, rather than dropping to a single point of failure while the fault is repaired. Multi-homed BGP lets us shift traffic between carriers in seconds and select the shortest path to each destination network rather than accepting whatever one provider advertises.
Our upstream networks, and their public PeeringDB records:
Interactive visualisation of our data centre locations, national long distance paths, and submarine cable connectivity across South Africa
Direct connectivity at South Africa's major internet exchanges for optimal routing and performance
AS329206Open peering policy. Port addresses and contacts are on PeeringDB.
Peer with RackzarDirect private peering with major cloud and content providers for optimal performance and reduced latency
Direct interconnect to Google Cloud Platform and Google services
Private peering with Azure and Microsoft 365 services
Direct Connect to AWS infrastructure and services
Private Network Interconnect with Cloudflare's global network
Direct interconnect to Oracle Cloud Infrastructure (OCI) and Oracle services
Our private interconnects give your servers high-speed, direct links to Google Cloud, Microsoft Azure, AWS, Cloudflare and Oracle Cloud. Traffic between your Rackzar infrastructure and your cloud workloads takes the shortest path, rather than crossing third-party transit networks.
Run databases and core systems on dedicated Rackzar hardware while front ends, storage or managed services live in the cloud, with fast links between them.
Combine Rackzar with one or more hyperscalers to avoid lock-in to a single provider and place each workload where it runs best.
Direct interconnects reduce hops between your environments, keeping API calls, replication and data transfers fast and predictable.
Replicate between Rackzar and the cloud over high-capacity links, so a secondary site is always close to your primary data.
If your organisation holds its own Autonomous System Number and IPv4 or IPv6 address space, you do not need to renumber when you move to Rackzar. We set up a BGP session with your ASN and route your prefixes across AS329206, so your own addresses work on our dedicated servers, colocation and private cloud.
Your prefixes are carried by our three upstream carriers and announced at the internet exchanges we peer at, giving your address space the same multi-homed resilience as our own network.
Our routing and switching runs on Juniper and Arista, platforms built for carrier networks rather than general-purpose x86 servers. It costs more to build this way, and it is why latency stays consistent when traffic peaks.
Routing
Juniper routers run the AS329206 backbone, carrying full BGP tables from all three upstreams and four exchanges with advanced traffic engineering.
Switching
A fully redundant Arista switching fabric connects every server, built for latency-sensitive applications and high-throughput workloads.
Forwarding happens in dedicated silicon, so latency does not drift with CPU load or traffic spikes.
Line-rate ports and redundant uplinks keep throughput level during busy periods instead of queuing and dropping packets.
MPLS, BFD, RPKI validation and full BGP tables run natively, with no software workarounds.
Strategic locations across South Africa's leading carrier-neutral data centres, including Uptime Institute Tier IV certified Africa Data Centres JHB2
Common questions about capacity, peering, redundancy and connectivity on AS329206
Our network peers at 100Gbit across major South African internet exchanges, providing extremely low latency and high throughput for all hosting services.
We peer directly at JINX, NAPAfrica, CINX and INX-ZA, ensuring optimal routing to ISPs, CDNs and cloud services across South Africa.
Three. We take IP transit from Workonline (AS37271), Agile (AS328748) and Network Platforms (AS37497), each over separate physical paths and separate BGP sessions. Losing any one carrier still leaves two fully redundant upstreams rather than a single point of failure.
Yes. We operate fully redundant fibre paths between facilities with automatic failover to ensure service continuity even during fibre cuts.
Yes. If you have your own ASN and IPv4 or IPv6 space, we can set up a BGP session with your ASN and route your prefixes across AS329206, so you can use your own addresses on our dedicated servers, colocation and private cloud. We need a Letter of Authority, valid RPKI ROAs and matching IRR route objects before we accept your announcements. See Bring your own IP space for details.
Yes. We maintain direct peering with major ISPs and local carriers for the lowest possible latency within South Africa.
Yes, private VLANs and cross-connects are available for enterprise clients needing direct connections between racks or facilities.
Our network is fully dual-stack with native IPv6 support across all platform layers including VPS, dedicated servers and cloud products.
We maintain POPs in Johannesburg and Cape Town through major carrier-neutral data centres to ensure low latency nationwide.
Whether you're interested in peering, colocation, or our cloud services, our team is ready to help you leverage our network infrastructure
