Data Center Archives - Page 5 of 11 - Fiber Optical Networking

25G WDM-PON for 5G Fronthaul Transmission

With the rapid development of mobile Internet, mobile data traffic and connected devices grow explosively. In response to the challenges brought by the network growth, 5G is under construction. 25G WDM-PON, combines the WDM technology and the PON topology, effectively bring high bandwidth, low latency, plug-and-play ONU (Optical Network Unit), simple OAM (Operation Administration and Maintenance) and low-cost advantages to 5G fronthaul transmission. Therefore, it has attracted widespread attention in the industry. This article describes 25G WDM-PON for 5G fronthaul transmission.

Technical Requirements for 5G Fronthaul Transmission

Compared to 4G architecture, the 5G RAN (Radio Access Network) architecture will evolve from the formal BBU (Base Band Unit) and RRU (Radio Remote Unit) two-level architecture of LTE to the three-level structure of CU (Centralization Unit), DU (Distribution Unit) and AAU (Active Antenna Unit). Therefore, the bearer network is also divided into front haul, middle haul, and backhaul. 5G base station fronthaul requires a bandwidth up to 25Gbps. However, as the number of base station increases, the cost of the base station to the fiber infrastructure networks and deployment will be higher. In this case, employing a Centralized Radio Access Network (C-RAN) architecture for 5G fronthaul holds great appeal for operators, but the current dark fiber solution for the C-RAN architecture requires a large number of mobile backhaul fibers. Most operators are still seeking an optimal solution to meet the needs of 5G fronthaul.

altTechnical Requirements for 5G Fronthaul Transmission

25G WDM-PON Solution for 5G Fronthaul Transmission

In the C-RAN architecture of 5G network fronthaul, the functionality of the 5G BBU will be reconstructed into two functional entities — CU and DU. The CU primarily includes the non-real-time part of wireless upper-layer protocol stacks and supports the distribution of some core network functions and the deployment of applications at the network edge. The DU mainly handles physical-layer functions and real-time transmission-layer functions. In order to save the transmission cost and reduce the transmission bandwidth between the RRU and the DU, some functions of the physical layer are moved down to the RRU implementation. 25G WDM-PON is a solution for transmission between DU and RRU. In the network architecture of WDM-PON 5G fronthaul, WDM-PON OLT and ONU are connected with DU and RRU respectively. The transparent business transmission between DU and RRU is realized by the adoption of the wavelength position multiplexing technology and AMCC (Auxiliary Management and Control Channel) technology. The OLT devices carry the midhaul service between DU and RRU at the same time that the OLT devices realize the fronthaul service between DU and RRU.

alt 25G WDM-PON Solution for 5G Fronthaul Transmission

Advantages of 25G WDM-PON for 5G Fronthaul

25G WDM-PON has advantages like high bandwidth, low latency, and low-cost. These all make 25G WDM-PON a better choice to satisfy 5G fronthaul requirements.

Reduce Capex

In the 5G construction, operators are faced with the pressure to reduce the number of sites and leased equipment rooms. In this case, employing a centralized radio access network (C-RAN) architecture can significantly reduce capital expenditure (Capex) on sites and equipment rooms. The 25G WDM-PON OLT can utilize an Access Office (AO) to deploy a centralized distributed unit (DU) pool. Thus, wireline and wireless AOs can be co-located if the condition permits.

Reduce Deployment Cost

The 25G WDM-PON 5G fronthaul solution allows for sharing of the existing fiber infrastructure, making it suitable for densely populated urban residential areas. The 5G network requires a large number of fiber resources. A network architecture based on the point-to-multipoint tree topology of a 25G WDM-PON can save a lot of fibers. The existing FTTx networks have rich lines and port resources in a wide range of deployments. 25G WDM-PON can completely reuse these resources to reduce 5G network deployment costs, avoid overlapping investments and improve 5G network coverage.

Conclusion

The 25G WDM-PON optoelectronic devices belong to the world’s leading edge, which requires high technology, chips, research and development capabilities. Overall, the 25G WDM-PON industry chain is basically mature, but it still needs continuous investment from all aspects of chips, transceiver modules (SFP, SFP+, SFP28), equipment and systems, to develop key technologies, reduce core device costs, establish a unified standard, and accelerate the productization process.

25GbE Ethernet and 25GbE Products FAQs

25GbE Ethernet has been an eye-catching topic ever since its introduction in 2016. After its appearance, the upgrade path to 100G has a new option — 10G-25G-100G. Although 25G has already been released, it’s still a new thing for many people. Here collect some common questions which will help you to better know 25GbE Ethernet and 25GbE Products.

25GbE Ethernet FAQs

What is 25GbE Ethernet?

25GbE Ethernet is a standard for Ethernet connectivity that benefits cloud and enterprise data center environments. It uses technology defined for 100GbE Ethernet, which implemented as four 25 Gbit/s lanes running on four fibers or copper pairs.

25GbE VS 10GbE, which is better?

As data centers require faster speed and higher bandwidth, 10GbE is not fast enough for switch-server connections. 25GbE was proposed to provide 2.5 times the performance of 10GbE as well as offers higher bandwidth at the same power consumption. 25GbE Ethernet supports 10GbE technology advancements in packaging and silicon. All these benefits contribute the existing switch architectures to supporting link speeds faster than 10GbE without the need to add cable/trace interconnects. At the same time, the architectures can keep up with the growth of the network bandwidth and become faster and richer.

25GbE VS 40GbE, who is the winner?

Compared to 40GbE, 25GbE can provide a higher port density with maximum switch I/O performance and fabric capability, enabling the network bandwidth to be effectively scaled in cloud and web-scale data center environments. What’s more, the 25G-100G (4x25G lanes) networking migration path provides a lower cost per unit of bandwidth by fully utilizing switch port capabilities when compared to 40G-100G upgrade path.

What makes 25GbE a choice for next-generation switch-server connections?

25GbE technology has two major advantages. First, by offering a single-lane 25Gbps variant, 25GbE allows for optimization of the bandwidth available from switch-fabric design, which is typically viewed from a per-pin and therefore per-lane standpoint. Another major advantage is that it takes existing module form factors and optical plants factors, such as SFP28 and QSFP28, and allows for a breakout connection that is configurable as either 25GbE per lane or the full 100GbE without changing the port on the front of switches and the wholes physical infrastructure.

Will the upgrade path of 100GbE be 10G-25G-100G?

Before the 25GbE specification was released, the 10G-40G-100G upgrade path is predominantly adopted by data centers to upgrade to 100GbE. However, due to the appearance of 25GbE, the 10G-25G-100G upgrade path has gained more recognition by offering the cost per bit, power consumption and server rack density advantages. 25GbE provides another option for 100GbE upgrade, and its prospect is bright.

25GbE Products FAQs

What kinds of 25GbE products are on the market today?

Currently, 25GbE products on the market are relatively complete. These 25GbE switch, 25GbE NIC, 25GbE SFP28, 25GbE DAC/AOC cable are all available on the market.

What does SFP28 stand for?

SFP28 is the enhanced version of SFP+, designed for 25G signal transmission. It has the same physical structure as SFP and SFP+, but its electrical interface is upgraded to handle 25Gbps per lane. The pinouts of SFP28 and SFP+ connectors are compatible. Compared to SFP+ solution, SFP28 has higher bandwidth, superior impedance control, and less crosstalk. In addition, the SFP28 copper cable has greater bandwidth and lower loss than the SFP+ version.

Can SFP28 cables be plugged into SFP+ ports?

Yes. SFP28 is backward compatible with SFP+ ports and can work sufficiently. SFP+ cables can be plugged into SFP28 ports but they are not designed for 25Gb/s data rates.

Can 10G SFP+ and 1G SFP transceivers be plugged into the associated 25G SFP ports?

Yes. The associated 25G SFP ports can run multiple speeds and support 10G SFP+ and 1G SFP transceivers. However, some port level restrictions may occur during speed mixing and matching.

Are branded 25G transceivers interoperable with third-party 25G transceivers?

Yes. Third-party 25G transceivers are compliant with relevant industry standard specifications and can be fully interoperable as long as compatibility is guaranteed.

Where can customers buy split cables for 100G to 4x25G connectivity?

Large vendors can supply volumes of MTP-LC multimode fiber breakout cables to customers for 100G to 4x25G connectivity. FS also provides 100G QSFP28 to 4x25G SFP28 DAC/AOC cables for you to choose.

Summary

25GbE Ethernet is considered an incremental update of 10GbE Ethernet, which supports 100GbE Ethernet with single lane at 25Gbps. Due to the booming of 25GbE, data centers will gradually turn to 25GbE networks. Knowing the above information is essential for you to deploy 25GbE Ethernet networks and avoid some problems when purchasing 25GbE products. FS offers a wide variety of high-quality and cost-effective 25GbE Ethernet portfolios such as 25gbe switch, 25G NIC, 25G SFP28 transceiver, and 25G DAC/AOC cables to meet your needs of 25GbE devices and connectivity options.

25G Server Adapter Selection Guide

25G server adapter, also known as 25G NIC (network interface card), is usually a separate adapter card that can be plugged into one of the server’s motherboard expansion slots. Since the server adapter establishes a network node, it must have a MAC address. The MAC address is burned into the server adapter at the factory, so it cannot be changed. So far, if you’re interested in 25G server adapter, you may wonder why and how to choose a suitable 25G server adapter for your network, here is the recommendation for you.

alt25G Server Adapter Selection Guide

Why Is There a Need for a 25G Server Adapter?

Storage space needs faster speed, especially when it comes to flash arrays. 10G is not fast enough, and many customers are looking for 25G, 50G, or even 100G capabilities. Public cloud builders are also looking to increase bandwidth and reduce link aggregation from their servers. It’s better and cheaper to have a single low-power adapter that has two 25 Gb/sec ports than to have four 10 Gb/sec ports on multiple adapters. That’s why the industry is very concerned about 25G products. The cost per bit and power consumption per bit of 25G products are lower.

What Is the Function of a 25G Server Adapter?

The main function of the 25G server adapter is to convert server data into electrical signals then send to the Ethernet switches, and vice versa. 25G server adapter is a kind of fiber-optic server adapter, usually used for fiber-optic networks. In most networks, fiber-optic server adapters are still too expensive for desktop use. So, they’re usually used for high-speed backbones. If a server connects to a high-speed fiber backbone, it will need a fiber-optic server adapter that matches the fiber-optic cable being used.

alt25G Server Adapter Selection Guide

How to Select the Best 25G Server Adapter for Your Network?

Whether a 25G server adapter can be properly selected, connected, and set up during networking is often the prerequisite for proper network connectivity. Here we name some mainstream 25G server adapters that are available on the market for your reference.

althttp://www.fiber-optical-networking.com/wp-content/uploads/2019/01/How-to-Select-the-Best-25G-Server-Adapter-for-Your-Network.png

As we see from the chart, those server adapters are almost the same. Most of them have a dual port SFP28, and just one of them has a single port SFP28. For users, it makes sense to have a dual port SFP28. In this case, the server adapter can handle more network traffic.

However, the most important factor lies in the interface PCIe. PCIe, shorts for Peripheral Component Interconnect Express, is a standard type of connection for internal devices in a computer. A PCIe based device physically slides into a PCIe slot on the motherboard. The PCIe interface allows high bandwidth communication between the device and the motherboard, as well as other hardware. And there are various versions of PCIe, for example PCIe 1.0, PCIe 2.0, PCIe 3.0, PCIe 4.0. Among them, PCIe 4.0 provides the fastest speed, supporting a bandwidth of 15.752 Gbit/s per lane (1969 MB/s). PCIe 3.0 is amazingly fast compared to PCIe 1.0 and PCIe 2.0, it can support a bandwidth of 7.877 Gbit/s per lane (984.625 MB/s). Since the server adapters mentioned above have the same version of PCIe, you can choose any one of them.

Selection Guide

Due to the limited number of 25G server adapters available on the market, when you are looking for one, keep in mind the matching products offered by the above companies. Generally, companies that provide more matching products are more reliable to your network. It can provide more options when expanding your network and prevent problems such as 25gbe switch and cables compatibility. If you are interested in FS 25G server adapter, feel free to contact us at FS.COM. We provide cost-effective 25G server adapters and a complete range of matching products just for your needs.

Open Network vs Closed Network: When and Which to Choose?

Internet plays an indispensable role in today’s society. From birth to now, it has continuously penetrated into various industries and has become the most important part of contemporary social technology and economic development. Therefore, Internet always gives people the characteristics of openness. But apart from the open network, there is the closed network that we rarely know. Let’s start by defining the two networks and make open network vs closed network to see which is more in line with people’s needs.
altOpen Network vs Closed Network: When and Which to Choose?
What Is an Open Network?

An open network can have many definitions depending on different requirements. An open network can refer to a network that must conform to open industry standards. Whether the device is used to interact within a network or externally, an architecture that is open every step of the way is needed. An open network can represent a network with an open ecosystem that defines the deployment scope of the solution set. An open network can mean an open source network that enables innovation in the marketplace while ensuring that the resulting product is constantly secured by the community that is contributing to it. An open network can indicate a network that provides access to the infrastructure in a programmable way through APIs (Application Programming Interfaces). However, those APIs are often closed out or only work with their solution. In summary, an open network allows a variety of entities to provide service on a reasonably equal basis versus each other and the network operator.

What Is a Closed Network?

Likewise, a closed network also has different meanings. A closed network can refer to a private telephone network that has no external (public switched telephone network) connectivity. A closed network can imply a network that uses proprietary technology which is not directly interoperable with other standards-based networks. A closed network can signify a WLAN that does not send its name (SSID) in beacon frames. Stations must know the SSID (Service Set Identifier) in order to connect to access points in that network. A closed network can represent a private network that can only be used by authorized devices. Outsider use is prohibited and enforced through cryptographic means. In short, a closed network is one that sets aside a great deal of the network capacity for a limited set of providers, usually but not always limited to the network provider.

Open Network vs Closed Network

From the above, we know how to define open and closed networks from different aspects. Here we list three different network environments. Take the forth definitions of open and closed networks as an example to compare and choose the network that works best for you.

Business Network

We know that all businesses depend on information flow. A closed network allows information to move freely within the business. And under the transparency provided, managers can directly know the situation and better guide the team’s work. The closed network also supports the need for privacy. Private messages can be sent to specific audiences, and private groups can be created for sensitive, ongoing conversations. Many closed networks provide “external network” capabilities that allow employees to connect outside partners, advisers, consultants, vendors, and suppliers. However, an open network allows employees to view more information, and when employees have the information they need, they make better decisions and are more productive. But in terms of privacy, an open network makes company information public and does not effectively protect the company’s core secrets.

Home Network

With the popularity of wireless networks, many houses are equipped with a wireless network. But is it okay to have an open home network? Not really. Having an open home network can be a security risk as it may allow anyone close enough to your router (e.g., a neighbor or war driver) to access your network and steal your personal information. In this case, a closed network will make your home network more secure.

Social Network

Essentially, social network is a broadcast distribution system designed to share as widely as possible. An open network means sharing with everyone and encourages open behavior. Participants simply “put it there” and less worried about individual reactions. Whereas a closed network means sharing is limited to selected people, interaction only happens between people who have specifically chosen to do so under a mutual trust. If the trust is broken, just rescind the interactive permission. The best social network depends on what the marketing professional is trying to achieve. To grab attention, fast and broad sharing open network is the best. If the idea is to create deeper engagement, then trust is more important, a closed network is more suitable.

Conclusion

Both open network and closed network provide unique value and can be used in different ways for the success of the network. They are widely used in network switches (e.g.,10gb switch). If you have figure out the difference between them, you can start building the network you need.
Related Articles:
FS Optical Transport Network Platform Unveiling
Open Network Linux: A Change Agent For Innovation

Open Network Linux: A Change Agent For Innovation

Driven by speed, economy, and the need for vendor freedom, many companies have been developing custom NOS (Network Operating System) on open networking platforms. To further advance network innovation, OCP (Open Compute Project) has been working to standardize open networking software and hardware. OCP’s Open Network Linux enables organizations to rapidly innovate and build modern, modular, next-generation NOS stacks by leveraging open source software on open network switches.

alt Open Network Linux, A Change Agent For Innovation

What Is Open Network Linux?

Open Network Linux is an OCP networking project. It is a Linux distribution for open hardware switches, namely network forwarding devices built from commodity components. It features an open networking hardware platforms with a unique architecture for pluggable forwarding stack and device management & programming software. It is the basic component for consumers who want to build NOS on top of open networking switches. Based on Open Network Linux, networking solutions can provide speed, freedom, and commercial efficiency and are now adopted by hundreds of data center organizations, service providers, mainstream enterprises, and SaaS/cloud providers. Consumers just need an ONIE (Open Network Install Environment) to install Open Network Linux onto on-board flash memory.

What’s the Competitiveness of Open Network Linux?

Over the years, many bare metal operating systems have emerged, such as Big Switch Network’s Switch Light, Cumulus Network’s Cumulus Linux, Broadcom’s Fast Path. Open Network Linux is a basic operating system that contains only example packet forwarding code. It is mainly expected that you or other projects will write their own packet forwarding code. Open Network Linux has two main goals. First, Open Network Linux aims to become a development platform for tinkers and DYI types to build their own forwarding applications. Second, Open Network Linux desires to be a useful component for building complete commercial solutions on top of bare metal switches. For example, the Switch Light OS of Big Switch Network is based on Open Network Linux. Therefore, the expectation is that people deploying Open Network Linux will deploy or build their own packet forwarding application on top. In addition, Open Network Linux expects that 3rd parties will provide binary-only forwarding applications (e.g., a traditional L2/L3 stack, an OpenFlow agent, etc.) over time. For example, the Open Route Cache or ‘orc’ code supports traditional L3 routing, and the Indigo binary works as an OpenFlow agent.

Supporters of Open Network Linux

There are many companies around the world that support Open Network Linux. For example, Finisar, Freescale, Accton, and Interface Masters provide platform drivers for Open Network Linux. Big Switch Networks and Open Networking add forwarding agents for Open Network Linux. And its hardware support companies include Edge-Core (Accton), Quanta, Dell, Mellanox, Netberg, Inventec, Celestica, HPE, DNI, Ingrasys, Alpha Networks, and a few unnamed/upcoming vendors. Open Network Linux is at the forefront of system support, as it continues to work with the partners and the community, it will further push the boundaries of innovation together.

Why Use Open Network Linux?

In the previous paragraphs we discussed so much about Open Network Linux, so why should we use Open Network Linux? There are several reasons to tell. Firstly, Open Network Linux helps the ecosystem focus on innovation. It helps to deal with many annoying software details to run an OCP switch and build platform drivers with common value asset. Secondly, Open Network Linux enables a reference NOS implementation. Hardware without software is not useful, it helps to package up details and best practices into one place. Thirdly, Open Network Linux help bootstrap the open ecosystem and OCP adoption. It allows commercial companies and DYI folks to build OCP based products faster.

Conclusion

Open Network Linux is a Linux distribution for bare metal switches (e.g., 10gbe ethernet switch, 100gbe ethernet switch, etc.). It supports OCP and non-OCP switches, ORC forwarding agent, and Indigo-based OpenFlow agent. For two years, we have witnessed open hardware and software have been shared with consumers, technologists, and vendors. As Open Network Linux continues to gain attention as a popular distribution for open network hardware, it will result in less integration work for hardware and software vendors, fewer bugs, and increased reliability as ONL based products are shipped to consumers.