As the demand for bandwidth in Cloud data centres is experiencing a notable surge, the networking and Ethernet industries are pivoting towards a fresh direction, namely 25G Ethernet. It appears that 25GbE holds greater favour and adoption among end users when juxtaposing upgrade paths of 10GbE-25GbE-100GbE and 10GbE-40GbE-100GbE. What prompts the selection of 25GbE? What advantages does it offer? This tutorial will comprehensively elucidate the merits of 25G Ethernet.
The Emergence of 25G Ethernet
Network engineers were once taken aback by the notion of a 10GbE link. However, virtualisation and cloud computing have introduced fresh networking challenges necessitating greater bandwidth. Top of Rack (ToR) switches, commonly boasting the highest number of connections in data centres, are swiftly surpassing the capabilities of 10GbE. Subsequently, the IEEE sanctioned standards for 40GbE and 100GbE to meet these demands. Nevertheless, 40GbE proves to be less cost-effective and power-efficient in ToR switching for cloud providers, while the deployment of 100GbE remains comparatively arduous and expensive.
Against such a backdrop, the 25G Ethernet standard was developed by the IEEE 802.3 Task Force P802.3by, designed for Ethernet connectivity to enhance Cloud and enterprise data centre environments. The 25GbE specification utilises single-lane 25Gbps Ethernet links and is founded on the IEEE 100GbE standard (802.3bj), achieving 100GbE through 4x25Gbps lanes.
25G Ethernet Optics & Cables
The 25GbE physical interface specification supports two main form factors—SFP28 (1×25 Gbps) and QSFP28 (4×25 Gbps). Commonly used transceivers are 25GbE SFP28.
The 25GbE PMDs (Physical Medium Dependent) specify low-cost, twinaxial copper cables, requiring only two twinaxial cable pairs for 25Gbps operation. Links based on copper twinaxial cables can connect servers to ToR switches, and serve as intra-rack connections between switches and routers. Fan-out cables (cables that connect to higher speeds and “fan-out” to multiple lower speed links) can connect to 10/25/40/50 Gbps speeds, and can now be accomplished on MMF (multimode fibre), SMF (single-mode fibre), and copper cables, matching reach-range to the specific application needs. Commonly used cables are 25GbE DAC and 25GbE AOC.
| Physical Layer | Name | Reach | Error Correction |
| Electrical Backplane | 25GBASE-KR | 1 m | BASE-R FEC or RS-FEC |
| Electrical Backplane | 25GBASE-KR-S | 1 m | BASE-R FEC or disabled |
| Direct Attach Copper | 25GBASE-CR-S | 3 m | BASE-R FEC or disabled |
| Direct Attach Copper | 25GBASE-CR | 5 m | BASE-R FEC or RS-FEC |
| Twisted Pair | 25GBASE-T | 30 m | N/A |
| MMF Optics | 25GBASE-SR | 70 m OM3 / 100 m OM4 | RS-FEC |
Why Choose 25G Ethernet?
While 10GbE is adequate for many current setups, it lacks the necessary bandwidth efficiency and demands extra devices, substantially raising costs. Additionally, 40GbE isn’t economically viable or power-efficient in ToR switching for Cloud providers. Hence, 25GbE was developed to overcome this dilemma.
Number of SerDes Lanes
SerDes (Serializer/Deserializer) is an integrated circuit or transceiver used in high-speed communications for converting serial data to parallel interfaces and vice versa. The transmitter section is a serial-to-parallel converter, and the receiver section is a parallel-to-serial converter. Currently, the SerDes rate is 25 Gbps. That means only one SerDes lane at 25Gbps is required to connect one end of a 25GbE card to the other end of a 25GbE card. In contrast, 40GbE requires four 10GbE SerDes lanes to establish connections. Consequently, communication between two 40GbE cards necessitates as many as four pairs of fibres. Furthermore, 25G Ethernet offers a straightforward upgrade path to 50G and 100G networks.
More Efficient 25GbE NIC for PCIe Lanes
At present, the mainstream Intel Xeon CPU only provides 40 lanes of PCIe (PCI Express) 3.0 with a single-lane bandwidth of about 8 Gbps. These PCIe lanes are used for not only communications between CPU and network interface cards (NIC), but also between RAID (Redundant Array of Inexpensive Disks) cards, GPU (graphics processing unit) cards, and all other peripheral cards. Therefore, it is necessary to increase the utilisation of limited PCIe lanes by NIC. A single 40GbE NIC needs at least one PCIe 3.0 x8 lane, so even if two 40GbE ports can run at full speeds at the same time, the actual lane bandwidth utilisation is only: 40G2 / (8G16) = 62.5%. On the contrary, a 25GbE NIC card only needs one PCIe 3.0 x8 lane, then the utilisation efficiency is 25G2 / (8G8) = 78%. Clearly, 25GbE is significantly more efficient and flexible than 40GbE in terms of the use of PCIe lanes.
Lower Cost of 25GbE Wiring
40GbE cards and switches utilise QSFP+ modules with relatively costly MTP/MPO cables not compatible with LC optical fibres of 10GbE. If upgrading to 40GbE based on 10GbE, most of the fibre optic cables will be abandoned and rewired, which can be a huge expense. In comparison, 25GbE cards and switches use SFP28 transceivers and are compatible with LC optical fibres of 10GbE due to a single-lane connection. If upgrading from 10GbE to 25GbE, rewiring can be avoided, which turns out to be time-saving and economical.
Distinct Benefits of 25GbE for Switch I/O
Firstly, 25G Ethernet has an excellent maximum switch I/O (Input/Output) performance and fabric capability. Web-scale and Cloud organisations can enjoy 2.5 times the network bandwidth performance of 10GbE. Delivered across a single lane, 25GbE also provides greater switch port density and network scalability. Secondly, 25GbE can reduce capital expenditures (CAPEX) and operating expenses (OPEX) by significantly cutting down the required number of switches and cables, along with the facility costs related to space, power, and cooling compared to 40GbE technology. Thirdly, 25GbE using a single lane 25Gbps Ethernet link protocol leverages the existing IEEE 100GbE standard which is implemented as four 25Gbps lanes running on four fibre or copper pairs.
Future 25G Ethernet Market Forecast
Over the past few years, 25G Ethernet has garnered escalating recognition, with 25GbE products undergoing notable advancements and securing an augmented market share. Anticipations hold that 25GbE will pursue a broader market in 2020 and will continue to flourish in the foreseeable future. Over the long term, 25GbE is forecasted to emerge as a future-proof trend in high-speed data centre networks, owing to the versatility of 25GbE adapters capable of operating at 10GbE speeds.
Additionally, 25GbE switches offer a more convenient pathway for migration to 100G or even 400G networks, circumventing the need for a 40GbE upgrade. Whilst the importance of industry consensus building cannot be overlooked, it is worth noting that currently, 25GbE is predominantly utilised for switch-to-server applications. Should the adoption of switch-to-switch applications be significantly encouraged, the potential for further advancement of 25G Ethernet is evident. In summary, the transition to 25GbE is gaining momentum.
Conclusion
Regardless of market research findings or user attitudes, 25G Ethernet appears to be the favoured choice moving forward, given its lower cost, reduced power consumption, and enhanced bandwidth. Considering the tangible advantages of 25G Ethernet, it is anticipated to continue advancing unquestionably in the future.
Related Articles：
How to Select the Optimal SFP-10G-SR Module for Diverse Networking Environments
