Are You Ready For 400G Ethernet?

The rapid development in telecom industry is driving massive demand for higher bandwidth and faster data rate, from 10G to 40G and 100G, will this keep going on? The answer is definitely “Yes”. Some time ago, migration from 10G to 40G or 25G to 100G has been a hot spot among data center managers. While recently, 400G solutions and 400G components are coming. Are you ready for 400G? This article will share some information about 400G Ethernet.

Overview of 400G

In the past couple of years, modules with four 25/28G lanes or wavelengths are the solutions for 100G Ethernet. However, they were expensive at the beginning. Until 2016, the optical components industry has responded to the demands with 100G solutions that already cost less per gigabit than equivalent 10G and 40G solutions, and new developments to further drive down cost and increase bandwidths. The next generation is 400G Ethernet. The IEEE has agreed on PSM4 with four parallel fibers for the 500 meters 400GBASE-DR4 specification that is part of the IEEE802.3bs standard being developed for approval by the end of 2017. The industry is already developing optical components for 400G Ethernet solutions. The following figure shows telecom and datacom adoption timelines.

Telecom and datacom adoption timelines

We can visually see that telecom/enterprise applications first adopted 100G technology in the form of CFP modules. Data centers generally did not adopt 100G interfaces until the technology matured and evolved towards denser, lower power interfaces, particularly in the form of QSFP28 modules. However, as the hyperscale data center market scales to keep pace with machine-to-machine communications needs, data center operators have become the first to demand transmission modules for data rates of 400G and beyond. Therefore, the 400G era is now upon us.

Modules for 400G

We know that the QSFP28 modules for 100G Ethernet and SFP28 modules for 25G Ethernet are now the dominant form factors. Though CFP, CFP2 and CFP4 modules remain important for some applications, they have been eclipsed by QSFP28 modules. To support higher bandwidth, what is the right module for 400G? The first CFP8 modules are already available. QSFP-DD is backward compatible with QSFP, and OSFP may deliver better performance, especially as networks move to 800G interfaces.

CFP8 module: CFP8 module is the newest form factor under development by members of the CFP multisource agreement (MSA). It is approximately the size of CFP2 module. As for bandwidth density, it respectively supports eight times and four times the bandwidth density of CFP and CFP2 module. The interface of CFP8 module has been generally specified to allow for 16 x 25 Gb/s and 8 x 50 Gb/s mode.

100G CFP to 400G CFP8

QSFP-DD module: QSFP-DD refers to Quad Small Form Factor Pluggable Double Density. It uses eight 25G lanes via NRZ modulation or eight 50G lanes via PAM4 modulation, which can support optical link of 200 Gbps or 400 Gbps aggregate. In addition, QSFP-DD module can enable up to 14.4 Tbps aggregate bandwidth in a single switch slot. As it is backwards compatible with QSFP modules, QSFP-DD provides flexibility for end users and system designers.

QSFP-DD vs QSFP

OSFP module: OSFP (Octal Small Form Factor Pluggable) with eight high speed electrical lanes is able to support 400G (8x50G). It is slightly wider and deeper than the QSFP but it still supports 36 OSFP ports per 1U front panel, enabling 14.4 Tbps per 1U. The OSFP is able to meet the projected thermal requirements for 800 Gbps optics when those systems and optics become available in the future.

OSFP module

Conclusion

Judging from the current trends, 400G will become the mainstream in the near future. But there are still some challenges for it to overcome, such as high capacity density, low power consumption, ever lower cost per bit, and reliable large-scale manufacturing capabilities. You never know what surprise the network will bring to you, let’s wait and see the 400G’s time.

How Far Can 25G Ethernet Go?

Seeing from the evolution of data transmission speed and size of data centers, it is not difficult to find that the pressure on the data centers to manage data quality and transmission speed continues to grow, which leads to the need for faster data transmission over the network. Ethernet industry has laid a path to higher networking speeds like 100GbE, and 25G Ethernet has been developed to provide a simpler path to future Ethernet speeds of 50 Gbps, 100 Gbps and beyond. The release of the 25GbE specification provides cost-effective solution for server-to-switch connectivity. However, network will not stop the pace of development. How far can 25G Ethernet go? This article is going to focus on the question.

Overview of 25G Ethernet

25G Ethernet is a standard for Ethernet connectivity in a datacenter environment, developed by IEEE 802.3 task force P802.3by. The IEEE 802.3by standard uses technology defined for 100 Gigabit Ethernet implemented as four 25 Gbps lanes (IEEE 802.3bj). In 2016, 25G Ethernet equipment was available on the market, such as 25G SFP28 transceiver and DAC cable. In addition, 25G Ethernet supports for 100G using QSFP ports that can be converted to 4 lanes of 25 Gbps, like 100G QSFP28 transceiver. Here is a table of 25G Ethernet specification for you.

25G Ethernet specification

Advantages of 25G Ethernet

For 10G ToR to 10G Server connectivity, the simplest cabling solution is to use two 10G SFP+ transceivers and one fiber optic cable. When the network has to be upgraded to 25G Ethernet, the data center manager only needs to replace 10G SFP+ transceivers with 25G SFP28 transceivers. In the same way, we know that in 40G ToR to 10G Server connectivity, one 40G QSFP+ transceiver, four 10G SFP+ transceivers and one MTP to LC breakout cable are utilized. When this network deployment is upgraded to 100G ToR to 25G Server connectivity, the work can be quickly finished by replacing 40G QSFP+ transceiver with 100G QSFP28 transceiver, four 10G SFP+ transceivers with four 25G SFP28 transceivers. It can be easily found that there are some advantages when upgrading from 10G to 25G or 40G to 100G:

  • It can offer both CapEx and OpEx savings through backward compatibility, for investment protection and seamless migrations with consistent rack-design and reuse of the existing cabling infrastructure, avoiding costly and complex changes.
  • The technology utilized in 100G to 25G connectivity is similar to that in 40G to 10G connectivity, but the performance is increased by 2.5 times, thus reducing the power and cost per gigabit significantly.
  • 25G Ethernet provides higher port and system density than a comparable 40G solution.
  • Both power savings and higher density results in lower cooling requirements and operational expenditure for data center operators.

advantages of 25G Ethernet

How Far Can 25G Ethernet Go?

Considering the significant benefits and compelling economics of 25G Ethernet, it is no surprise that the move to 25G is accelerating—a recent five-year forecast by industry analysts at the Dell’Oro Group predicts 25G Ethernet will be the dominant Server port speed for new systems by 2018. You can learn about it from the following figure.

forecast for 25G Ethernet

However, never underestimate the need for industry consensus building. At present, 25G is mainly used for switch-to-server applications. If it can realize switch-to-switch application, 25G Ethernet may go further.

Conclusion

To be frank, 25G Ethernet indeed gains ground in some aspects compared to 10G and 40G Ethernet. If you plan to deploy 25GbE network, you can visit FS.COM which provides quality 25G SFP28 transceiver and various fiber optic cables.

Related articles:

Taking an In-depth Look at 25G SFP28

25G Vs. 40G Ethernet: Who Is the Winner?

25G SFP28 Cable: The Most Economical Option for ToR Server Connection

The Changing Network Architecture in 25G/100G Data Center

25G Vs. 40G Ethernet: Who Is the Winner?

In recent years, the fast growth of data centers leads to increase in global data traffic, which give rise to the need for faster data transmission over a network. 25G Ethernet is the product of that condition. 25G Ethernet is regarded as an incremental update from 10G Ethernet, and it supports 100G Ethernet with single lane at 25Gbps. Due to the booming of 25G, some industry experts claimed that 40G Ethernet is dead, which is biased in some degree. Then 25G vs. 40G Ethernet, which is even better?

Advantages and Disadvantages of 25G and 40G Ethernet
Advantages of 25G Ethernet

Different from 40G and 100G, 25G is a single-lane variant for 25Gbps operation, and that allows a breakout of 100G, which fit the most popular form factors. Based on existing module form factors, such as SFP28 and QSFP28, 25G operations allow for a breakout connection that is configurable as either 25G per lane or the full 100G without changing the port on front of switches, bring more flexibility in the rack and front-panel connections. In addition, with the 25G Ethernet, network operators are no longer forced to use a 40G QSFP port to go from one individual device to another to achieve 100G throughput. The final advantage of 25G is that it can use existing optical plants (depending on what was installed) and increase the bandwidth by 2.5x without changing the physical infrastructure.

25g ethernet sfp28

Disadvantages of 25G Ethernet

As 25G Ethernet is just rising, the interoperability becomes an important factor to ensure wide market adoption and to offer higher speeds for future applications. Besides, compared with 10G 40G and 100G, there aren’t many products of 25G, which limit the development of 25G networks.

Advantages of 40G Ethernet

40G Ethernet is an Ethernet standard developed by the IEEE 802.3ba Task Force to support sending Ethernet frames at 40 gigabits per second. It also addresses physical layer specifications for communication across back planes, copper cabling, multimode fiber optic cable, and single mode fiber. And 40G Ethernet technology is more mature compared with 25G. In the market, there are various types of products for 40G applications, especially the MPO trunk cable assemblies, cassettes, and QSFP 40G optical modules, which offers the required bandwidth for different applications.

40G solution

Disadvantages of 40G Ethernet

At present, 40G is popular in data centers and no drawbacks found. If we have to say one, the utilization of fibers may be one. As we all know, 12-fiber cabling solution is common in 40G networks. But there are four fibers unused, resulting in fiber waste.

Comparison Between 25G and 40G Ethernet in Network
Application

At present, 25G is mainly used for switch-to-server applications. While 40G is for switch-to-switch applications. In other words, no one is using 25G for switch-to-switch links right now. Even the industry giant like Cisco doesn’t offer 25G optical transceiver. But with the fast development of 25G Ethernet, 25G for switch-to-switch application maybe come into reality in the near future.

Switches Selection

Switches are important when comparing 25G and 40G Ethernet. Most switches are currently sold, like Cisco 93180YC-EX, Arista 7060CX-32S support both 10G and 25G, and the price is not higher than older 10G products with full backward compatibility. For example, each SFP28 port supports 1G, 10G or 25G, and each QSFP28 port supports 10G, 25G, 40G, 50G or 100G.

Cabling Options

Cabling options determine how far the two types of Ethernet go. It’s a big mistake to ignore cabling. In the market, there are several types of cabling options, and there are some big swings in price. Here is a simple comparison.

25g ethernet vs 40g ethernet

Summary

From the comparison above, we can draw a conclusion that 25G Ethernet can be used for data centers, but it doesn’t mean 40G is dead. Even though 25G Ethernet seems to have a brilliant future, under present conditions, 40G is a safe choice due to its mature market adoption. Perhaps in a few years, 25G connectivity will be a cheaper alternative.

Related articles:

Taking an In-depth Look at 25G SFP28

How Far Can 25G Ethernet Go?

25G SFP28 Cable: The Most Economical Option for ToR Server Connection

The Changing Network Architecture in 25G/100G Data Center