Ethernet high capacity

Connect centres when data volumes demand more capacity.

High-capacity Ethernet retains a Layer 2 (L2) interface. It suits replication and backups; measure volume, window and concurrent traffic before sizing.

Ethernet simplifies the interface but does not guarantee a diverse route or protection on its own.

Let’s talk
Racks and optical fibre in a data centre setting.

How it works

Size data-centre connectivity around the data it must carry.

Data centre A
Layer 2 Ethernet transport
Replication and data
Data centre B

From data volume to circuit design

Start with volume, transfer window and concurrent traffic. Layer 2 handover connects data-centre equipment; routing and security remain separately defined.

Validate port speed, committed capacity, VLAN and MTU. For replication, add storage-platform latency and protocol constraints.

Data centre A
Layer 2 Ethernet transport
Bursts and backup windows
Data centre B

When backup misses its window

Compare offered load and effective transfer, errors, congestion and server limits. Nominal port speed is not application throughput.

When backup and replication overlap, adjust windows and agreed priorities before upgrading. Define the metric that will demonstrate improvement.

Data centre A
Layer 2 Ethernet transport
Replication continuity
Data centre B

Network continuity and data consistency

An alternative circuit must support essential load and avoid shared failure domains. The application determines how it handles disconnection, delay or queued data.

Test failure, recovery and resynchronisation. A restored link is not the same as a fully consistent replica.

Illustrative case: from situation to decision

Situation
The overnight backup does not finish before morning activity begins.
Option
Calculate useful throughput from volume, window and concurrent flows; check storage too.
Check
Measure a representative backup and replication consistency, not just port speed.

More Ethernet capacity does not replace replication design, storage or application recovery.

Decisions for your project

Demand
Volume, windows, concurrency and expected growth.
Endpoints
MTU, VLAN, ports and actual server/storage capacity.
Recovery
Alternative route, essential service and consistency validation.

From design to operation

Measure representative load, size the circuit and accept it using transfer and recovery tests.

Complements, alternatives and variants

Product sources

Back to top