E-Line is the point-to-point service type in Carrier Ethernet, as defined by MEF, the industry body that standardizes it. A provider connects two of your sites with a private Ethernet circuit: you plug into an Ethernet port at site A and another at site B, and traffic passes between them at Layer 2 under the provider’s service level agreement. E-Line comes in two main forms, the port-based Ethernet Private Line (EPL) and the VLAN-aware Ethernet Virtual Private Line (EVPL). Buyers often think of it as an Ethernet leased line.
At a glance
- E-Line connects exactly two points; for multipoint connectivity the equivalent is E-LAN.
- EPL gives one circuit per port with high transparency; EVPL lets several circuits share one port using VLAN tags.
- Performance is defined by a bandwidth profile and, often, classes of service, with targets for delay, jitter and loss that vary by provider.
- It is private Layer 2 connectivity, not internet access; you handle IP routing between the sites.
- Price and lead time depend on distance, whether sites are on the provider’s network, speed, diversity and term.
What problem it solves
Organizations often need a dependable, private link between two locations: a primary and backup data center, a headquarters and a large branch, two campus buildings, or an office and a colocation site where cloud on-ramps live. Running that traffic over the internet means variable latency and jitter and no performance commitment; building private fiber is expensive and slow.
E-Line gives you a provider-run circuit that behaves like a long Ethernet cable between the two sites, with a defined bandwidth profile and contracted performance targets. Because it hands off as standard Ethernet, it plugs straight into existing routers and switches.
How it works
Handoffs. At each site the provider delivers an Ethernet port, called a user network interface (UNI), at a speed such as 1, 10 or 100 Gbps, over copper or fiber. This is where the provider’s responsibility ends and where performance is measured.
The virtual connection. A logical circuit, called an Ethernet virtual connection, links the two ports across the provider’s network. Underneath, providers may use MPLS, Ethernet switching or optical transport; you see only Ethernet at each end. Traffic is kept logically separate from other customers.
EPL or EVPL. With EPL, one circuit uses the whole port and the provider passes your traffic, including your own VLAN tags, largely untouched. With EVPL, the port is VLAN-aware: the provider maps your traffic to circuits by VLAN, so one port at a hub can carry circuits to many sites.
Bandwidth profile. The service has a committed information rate (CIR), the rate the provider commits to carry, and on some services an excess rate for bursts that may be dropped when the network is busy. Many providers offer several classes of service, for example one for real-time traffic and one for best effort.
Monitoring and testing. Providers commonly use Ethernet operations and maintenance standards (such as IEEE 802.1ag and ITU-T Y.1731) to check continuity and measure delay and loss. At turn-up, activation tests such as ITU-T Y.1564 or RFC 2544 confirm throughput, latency and frame loss.
When it matters for buyers
- Data center and site interconnection. Storage replication, backups and clustered systems need predictable performance between two points.
- Choosing between connectivity types. E-Line suits a few important point-to-point links. MPLS or SD-WAN over dedicated internet access often suits many sites better, and a wavelength service may suit very high capacity.
- Resilience planning. Two E-Lines add little resilience if they share a physical route. Ask for route documentation; path diversity claims need checking.
- Reaching cloud and colocation. E-Line is a common way to reach a colocation facility where cloud on-ramps are available.
Our Private Networking guide compares E-Line with other private connectivity options.
Questions to ask vendors
- Is this EPL or EVPL, and which of our VLAN tags and Layer 2 control frames will you pass through or filter?
- What is the committed rate, is there a burst or excess rate, and what happens to traffic above the commitment?
- Which classes of service are available, and what are the delay, jitter and frame loss targets for each?
- What is the maximum frame size (MTU) supported end to end?
- Are both sites on your network today, and what is the expected installation interval?
- If we buy two circuits for redundancy, can you document that they follow physically separate routes?
- What availability and repair-time commitments apply, and what credits do we get if you miss them?
- Will you share turn-up test results and ongoing performance reports?
How it differs from E-LAN
E-Line connects exactly two points. E-LAN is the multipoint Carrier Ethernet service type: it connects several sites so each can reach the others directly, as if they were on one shared LAN, and new sites can join the same service. Many networks use both, with E-LAN between main offices and E-Line for specific point-to-point links such as data center replication.
Problems this helps with
- Connecting two data centers or offices privately without building our own fiber
- Replication and voice traffic suffering on internet-based links
- Not knowing whether to buy EPL, EVPL, MPLS or a wavelength
- Circuits sold as "diverse" that share the same route
