What is a Core Network?

A core network is the main infrastructure for routing network traffic and is typically owned by a telecommunications provider or internet service provider. A core network connects and routes traffic between access networks and edge networks and to other core networks, supporting mobile networks and cellular connectivity. The core network handles large amounts of data from multiple access networks, providing the backbone for mobile, broadband, and cloud-native architectures that need network connectivity.

A core network, also known as a backbone network, has several essential functions that support data communications.

Traffic routing and switching

A core network routes and switches packets between edge nodes, access networks, and core domains. It is responsible for determining the most efficient path that data can take to its destination, allowing telecommunications data to flow to its destination. The packet-switched core network differs from the public switched telephone network, which is used almost exclusively for telephone calls.

Routing refers to deciding where data packets should go, while switching manages how these packets actually move through the network.

A core network uses routing protocols such as the Interior Gateway Protocol (IGP), Intermediate to Intermediate System (IS-IS), and the Open Shortest Path First (OSPF) to route within a single network. For more complex movements between networks, a core network can also use Exterior Gateway Protocols (EGPs) such as the Border Gateway Protocol (BGP).

For switching, a core network typically uses Multiprotocol Label Switching (MPLS), which forwards packets with short, fixed-length labels. Instead of having to spend time performing IP lookups every time the packet hops, this strategy allows data packets to pass through the system more quickly.

Load balancing

Load balancing systems aim to distribute network traffic across all available routes, making sure no one pathway is overloaded with packets. By equally distributing all incoming traffic, routing has a more predictable and consistent speed. In a core network, load balancing is mainly done through using Equal-Cost Multi-Path (ECMP) routing and policy-based load balancing defined by network administrators.

Network resilience and failover

Due to how essential telecom networks are, core networks have built-in resilience to prevent downtime in the case of a primary node failure. If a pathway, fiber link, or router in the network goes offline, physical redundancy with alternative possible routes means that data continues to flow. Efficient network segmentation technology helps to reduce the impact of one part of the network shutting down.

Systems administrators can configure topology convergence, allowing routing protocols to recalculate a new path instantly if they encounter a failure in their intended path.

Quality of Service (QoS) management

A vast array of different telecommunications data can flow across a core network, from live video streams to voice calls. QoS management includes any mechanism that service providers use to make sure that any important data gains priority over other data packets. For example, you could mark certain high-priority data with a specific MPLS label to accelerate its movement through the system.

Network security and access control

Core networks employ a range of cybersecurity and network defense tools and systems to keep them safe from unauthorized access and policy noncompliance. Most commonly, core networks have defenses such as intrusion detection, access control, route filters to ignore potentially malicious routes, and control-plane policing to protect from disruptive bot traffic. Network connections are also encrypted to protect data in transit and at rest and to stop unauthorized parties from accessing this data.

A core network has four main segments that work together to move routing traffic through its network architecture.

Ingestion and entry points

Data traffic enters the core network through an entry point, most commonly by an ingress router or a Provider Edge (PE) router, which aggregates traffic from mobile networks, cloud native applications, and other networks. These ingress points sit between the core network and access networks (where an end user interacts) and handle all incoming data.

Once data arrives at the core, it is aggregated and classified. If necessary, protocols such as the MPLS attach additional labels to a data packet. If there are any QoS policies to apply, this is the moment where the system adds that metadata to the data packets.

Routing decisions with the control plane

The control plane is the part of the core network that manages the selection and building of the routes a network uses to transfer data. When the control plane receives new data packets through ingestion, it will use routing protocols, like OSPF, IS-IS, and BGP, to exchange topology information across all available routers. Routers receive this information in their Routing Information Base (RIB), using it to calculate the best possible path through the network.

The best path can depend on a range of factors, such as if certain routers have additional administrative rules in place that might block some forms of data or the current link metrics of individual routers. Artificial intelligence is increasingly playing a role in routing decisions for low latency and cost efficiency.

After selecting a route, the packet installs that information in the Forwarding Information Base (FIB). The data plane then uses this information to begin forwarding the data packets. As the data plane works independently of the control plane, they both handle their tasks in isolation, helping to provide stability and scalability.

Traffic forwarding

The data plane is fully responsible for the actual movement of data through the system and across router hops. Each router’s data plane has hardware such as Application-Specific Integrated Circuits (ASICs) that allow it to forward data packets at line rate. Line rate is the theoretical maximum rate of data transmission, with ASICs or Network Processing Units (NPUs) allowing routers to get close to this speed.

When data arrives at a router on its journey, the router checks its destination, either in the FIB or on its specific label, if traveling by MPLS. The router then makes any necessary changes, such as updating the delivery time, before transmitting back out by the egress interface to the next router. This router-to-router journey is known as hopping.

Exit points and handoff

As data moves through the core network toward its final destination, it will reach an egress point where it wants to leave. These are often PEs or border routers that act as a bridge between the core network and the next network domain that data wants to pass to.

The egress point routers will prepare data to leave the core, delivering it either to an access network, another provider’s core, or a cloud service network. Egress routers also remove any MPLS labels, as the data has successfully traveled across its core and won’t need these anymore.

A BGP manages data exchange here, allowing the data to pass on to its next location by either Private Network Interconnects (PNIs) or Internet Exchange Points (IXPs). After entering a new edge or access network, the data packets move through local nodes to the end user or application server where they were originally destined.

Here are some of the core network technologies used in a core network and in communications data routing.

MPLS (Multiprotocol Label Switching)

MPLS is an alternative to IP routing that uses short labels to add metadata to packets. These labels dictate how certain packets will move through the core, such as whether they need priority or if they need to take specific routes. MPLS also enables faster movement through the core, as label lookups are faster than full IP calculations at every hop.

BGP (Border Gateway Protocol)

BGP is a routing protocol that handles how data moves between large networks. Based on its configurations or any policies that apply, BGP decides the best pathway for data to move across the Internet or between different network providers. In a core network, BGP is essential, as this protocol makes sure all routes consistently apply routing rules. It also helps different network carriers communicate as data leaves one system and enters the next.

Core networks typically use internal BGP (iBGP), whereas between network links use external BGP (eBGP).

Optical transport networks

Optical transport networks (OTNs) are standard digital transport layers that carry data as light. OTNs can send several signals on one single fiber, using Dense Wavelength Division Multiplexing (DWDM) to facilitate massive data movement. These networks are a core technology of the network design that allows internet and cloud services to transfer data at scale.

Software-defined networking (SDN)

A Software Defined Network (SDN) is what separates the control plane from the data plane. This virtualized separation allows administrators to optimize performance and use network automation to perform certain tasks. You can manage SDNs through centralized software controllers.

There are key differences between core networks, access networks, and edge networks.

Access network

The access network is an outer layer in network infrastructure that acts as a bridge between end users and the central core networks. Incoming and outgoing data will move through the access layer.

Core network

The term core network refers to the central backbone of network systems, providing much of the underlying infrastructure that facilitates data transfer and routing at scale. For mobile communications, the evolved packet core (EPC) architecture in the core handles 4G and 5G delivery between service provider networks. A core network can sit between distributed systems.

Edge network

An edge network helps to provide low-latency data transfer by providing compute and storage services relatively closer to the user. For real-time video or cloud apps, edge nodes are able to preprocess data to reduce demand on the system and improve speed.

AWS owns and operates our own core networking to help customers gain access to their cloud-based resources fast. AWS offers various services that can help with improved connectivity between your own network resources and links to and within the AWS cloud:

• AWS Cloud WAN provides a central dashboard for making connections between your branch offices, data centers, and Amazon Virtual Private Clouds to build a global network.
• AWS Direct Connect allows you to create a dedicated network connection to AWS. With AWS Direct Connect, while in transit, your network traffic remains on the AWS global network and never touches the public internet.
• AWS Transit Gateway is where you connect Amazon VPCs, AWS accounts, and on-premises networks to a single gateway. AWS Transit Gateway acts as a cloud router to scale your networks.
• Amazon Virtual Private Cloud (VPC) allows you to define and launch new services and AWS resources in a fully logically isolated virtual network on AWS.

Get started with core network architecture on AWS by creating a free account today.