Author：Takfly Group Date：2019-09-24 Tags：25G
With the increasing demand for high-bandwidth in private cloud, public cloud data centers and service providers, 25G and 100G are widely used, and 200G/400G optical devices will be produced and shipped in 2019. So far, most server vendors have begun to offer 25G fiber NICs as I/O (input/output) options for servers. The signal transmission rate of Ethernet has increased from the previous 10G to 25G, 100G and even higher. Although 1G, 10G and 40G currently occupy a major share of the Ethernet port market, future demand for 25G and 100G will be stronger than ever, because high bandwidth is undeniably driving the data center toward higher The direction of scalability and flexibility is growing.
1. Why is the data center ushered in 25G?
The data center is expanding at an unprecedented speed, driving the need for higher bandwidth between the server and the switch. To meet this trend, the access network has gradually upgraded from 10G to 25G, which is also the connection between the server and the ToR switch. Provides high density, low cost and low power solutions.
25G development history
Since the introduction of 25G in 2014, companies such as Google, Microsoft, Arista Networks, Broadcom and Mellanox have been pushing the development of the 25G Ethernet standard to achieve a 25G top-of-rack server network. As the heat of 25G continues to rise, 25G will spread rapidly in the market, and the future 25G will provide a comprehensive solution for the connection between the server and the switch. The chart below shows the key milestones for 25G Ethernet from 2014 to 2018.
2. Talk about the advantages of 25G from the existing network upgrade method
Before the release of the 25G Ethernet standard, the network upgrade mode of data centers such as enterprises and operators is generally 10G-40G, accompanied by the official release of the 25G Ethernet standard, and the 25G-100G network upgrade method with low cost and low power consumption. And the advantages of high density have gained more applications and promoted the rapid development of 100G Ethernet. Let's take a look at what is the difference between 10G/25G/40G and which upgrade method is superior.
25G offers higher performance bandwidth than 10G
In current data centers, the network connection between the server and the switch typically chooses between 10G and 25G. Compared to 10G, 25G is based on 10G package and chip technology, providing higher bandwidth and performance. The emergence of 25G enables the data center to be based on the existing network architecture, without adding any cable interconnections, and can support higher speed (over 10G) transmission, meeting the needs of future networks for higher bandwidth, and upgrading the network. More convenient and easy. The wiring infrastructure required for 25G and 10G transmission is basically the same, which can effectively avoid the cost and re-wiring complexity, making network upgrade more convenient.
In addition, 25G is similar to 10G, using SerDes (Serlializer and Deserializer, the abbreviation of English serial deserializer) single channel, with backward compatibility, significantly reducing power consumption and cost, helping data center operators save capital expenditures And operating expenses.
What kind of speed will we get when we plug a 25G SFP28 optical module into a 10G SFP+ port?
In theory, the 25G SFP28 optical module is backward compatible with 10G SFP+ ports at speeds up to 10Gb/s. However, this method of use is not suitable for all brand switches and optical modules, considering the limitations of fiber-optic network cards and switch ports. This method is generally not recommended.
25G is more in line with high density requirements than 40G
The 25G interoperates with the Ethernet specification between the server network adapter (NIC) and the switch, providing greater port density and lower unit cost for the rack server's bandwidth connection. Compared to 40G, it provides higher port density and maximizes switch I/O performance and switching capabilities. From the point of view of the number of connected switches, since 25G uses single-channel technology (only one 25G port is occupied), and 40G uses four channels (occupies four 10G ports), there are more switches that can be connected to 25G. In addition, compared to the existing 10G and 40G, the transmission performance of 25G and 50G is improved by 2.5 times.
For large and high-end enterprises, the port density of the server largely determines the cost of the cabling and switch infrastructure in the entire system. Therefore, the cost of upgrading from 25G to 100G is relatively low compared to 40G. Because the switch port is fully utilized when the 25G-100G network is upgraded, the bandwidth cost is effectively reduced.
Has it been upgraded to a 40G network, is it necessary to deploy a 25G network?
Compared with 40G equipment, 25G equipment is more expensive, so it has been upgraded to 40G network. Is it necessary to deploy 25G network? Due to the rationalization of the cost of the 25G channel, the future 25G is definitely an important path for 10G upgrade to 100G or higher. If you need to increase the baud rate (signal transmission rate) or plan to upgrade the network to a higher speed (100G/200G/400G), you must deploy a 25G network. If there is no demand, then you do not need to deploy a 25G network.
Looking at the prospect of 25G from the future 100G/200G/400G data center
At present, 25G servers and 100G switches have been seen everywhere in ultra-large-scale data centers. They have gradually replaced the previous 10G servers and 40G switches. This network upgrade has increased the throughput of the entire system by 2.5 times and reduced the number of channels. Equipment incremental cost. As the Ethernet industry continues to innovate and lay the foundation for higher-speed R&D, the 25G-100G upgrade model has become an important path for data centers.
25G offers more possibilities for 50G
As we all know, 25G requires only one channel to provide 2.5 times bandwidth compared to 10G. In the future, 50G can provide 1.25 times bandwidth with one channel compared to 40G. Currently 50G has been proposed as 100G/200G/ The basis for the 400G network upgrade, but the implementation of the 50G Ethernet standard will take some time.
25G will provide more possibilities for 50G. Since the implementation of 50G Ethernet can be based on two 25G channels, it will become an alternative to using 40G for four 10G channels, reducing the cost of network equipment in the data center by reducing channel redirection. . In the future, the network upgrade path may evolve from the traditional 10G-40G-100G to 10G-25G-50G-100G. In any case, upgrading a data center to a 50G or 100G network over multiple 25G channels will be simpler and more economical.
25G lays the foundation for 200G/400G network upgrade
The 25G/50G/100G network architecture is more flexible and is often used as a solution for large data centers, paving the way for later 200G/400G upgrades. High-end enterprises and large data centers are now shifting from this, effectively driving the interconnection between implementations within large data centers and data centers. More and more suppliers in the market are now working on the development of 200G/400G optics, some of which have been successfully put into use. The implementation of 100G Ethernet is based on the development of 25G/50G. Similarly, the future 200G/400G network upgrade will be based on 100G. The table below lists the paths to upgrade to 200G/400G based on 25G/50G/100G.
If you have any question about 25G SFP28 optical transceiver module,please contact with email: email@example.com
Please click below link to leave messages,
we will give you an active reply ASAP.