A Glimpse Into The Future: 25G & 50G Ethernet

With the ever growing usage of 10G network, 10G could not satisfy the requirement for some Ethernet network users who urge for a higher demand on speed, distance, media and cost. Under this circumstance, upgrading network is paramount. For 100G network upgrading, there are three available approaches, “10G—40G—100G”, “10G—25G—100G” or “10G—25G—50G—100G”. The latter two are announced to better satisfy the data center and cloud network. Comparing to 40G and 100G, people heard less about 25G and 50G. So what are they? This article would put emphasis on 25G Ethernet and 50G Ethernet as well as their optics.

25G-100G immigration

25G Ethernet

25 Gigabit Ethernet, abbreviating as 25G Ethernet, is standard for Ethernet network connectivity. Developed by IEEE P802.3by 25 Gb/s Ethernet Task Force, 25G Ethernet is a standard for Ethernet connectivity. The 25 Gigabit Ethernet Consortium is an open organization to all third parties who wish to participate as members to enable the transmission of Ethernet frames at 25 or 50 Gigabit per second (Gbps) and to promote the standardization and improvement of the interfaces for applicable products. The main features of 25G Ethernet are listed in the following:

  • A single lane per physical port maximizes the number of connected servers or uplinks per switch.
  • Single higher speed 25 Gb/s lanes maximize bandwidth and switch fabric utilization vs. 4 x 10 Gb/s lanes.
  • Overall higher port count, utilization and total server interconnect bandwidth vs. 40 GE.
  • Connections to switch ASICs is limited by SERDES count and bandwidth.

SFP28 Pluggable Modules

SFP28 is the abbreviation of Small Form-Factor Pluggable 28, which is the third generation of SFP interconnect systems. The SFP28 optical module is designed for 25G performance and developed by the IEEE 802.3by specification. According to the SFP28 Multi-Source Agreement (MSA) and SFP28 specification, the SFP28 is designed with a form factor, optical/electrical connection and digital diagnostic interface. In addition, the SFP28 optical transceiver has also been designed to meet the harshest external operating conditions including temperature, humidity and EMI interference. Below are the industry standard 25G optics:

industry standard 25G optics

50G Ethernet

Comparing to 40G Ethernet, 50G Ethernet is more rarely known by people. Being led by the 25G Ethernet Consortium, 50G Ethernet is initially based on 2 lanes of 25 Gb/s. IEEE802.3bs is the 50G per lane specifications to support Nx50G configurations. And the standard expected in September 2018 while the interface expected on the market in 2018+. Different from 40G Ethernet, 50G initial limited deployment as proprietary 2x25G. In terms of technology, 40G and 50G per lane (Serial) technology will be defined together (40G as reduced speed 50G). With the respect of cost, 40G and 50G Serial will have similar cost, i.e. 50G Serial will offer 25% more bandwidth for the same cost. The core features of 50G Ethernet are listed in the below:

  • A faster base signaling rate is needed to for higher capacity.
  • Similar to 25 GE, 50 GE extends existing common network topology for higher speed.
  • The server and data center market requirements vary widely.

50G Pluggable Modules

New 50 GE pluggable modules are in the same common form factor sizes as other common pluggable modules. There are two form factors of 50G modules, SFP56 and QSFP56. The SFP56 pluggable module has the same size as SFP, SFP+ and SFP28 while the QSFP56 pluggable module has the same size as QSFP, QSFP+ and QSFP28.

Conclusion

Through this article, we are cleared the 25G and 50G Ethernet as well as their optics respectively. With the ever increasing usage of network data due to millions of new connected devices to servers and storages data centers, 25G Ethernet and 50G Ethernet provide a flexibility, scalibilty, cost-efficient way for adapting to future network growth.

10G Technology: 10GBASE-T Vs. SFP+

As the basis of upgrading network, 10G network has been ubiquitous in data center, enterprise network and even home networking. 10GBASE-T and 10G SFP+ are two different kinds of technology which transmit data via copper and fiber respectively. 10GBASE-T technology provides the most flexible and economical solution while 10G SFP+ offers the compatible and user-friendly solution for 10G Ethernet connectivity option. This article would shed light on the difference between 10GBASE-T and 10G SFP+.

10GBASE-T Technology

As the fourth generation of IEEE standardized Base-T technologies, 10GBASE-T is designed to reduce overall costs and improve flexibility. By using RJ45 connectors and unshielded twisted pair cabling, 10GBASE-T allows 10Mbps, 100Mbps, 1Gbps, and 10Gbps data transmission, while being backward-compatible with prior generations. Merits and demerits of using 10GBASE-T are listed in the below.

Pros of 10GBASE-T
  • Cheap twisted pair cables.
  • Patch panels can be used without messing around with transceivers.
Cons of 10GBASE-T
  • Higher power consumption.
  • People may get tempted to use substandard cabling, and this would have a negative influence on the speed.
  • No good way to extend length beyond 100m (though this can be somewhat mitigated by choosing switches with mostly 10GBASE-T but also a handful of SFP+ ports) limited choice of equipment.
10G SFP+ Technology

The 10G SFP+ transceiver meets the standard of Multi-Sourcing Agreement (MSA), and provides the cost effective solution for 10G optical data communication. It supports both duplex and simplex LC optics interfaces. The 10G SFP+ transceiver consists of 10Gbit/s DFB/EML optical transmitter and PIN receiver, which allow 300m~120km 10G Ethernet and 10G fiber channel applications. Advantages and disadvantages of using 10G SFP+ transceivers are listed in the below.

Pros of SFP+
  • Lower latency
  • Lower power consumption
  • Cheaper NICs and switches
  • More choice of connected equipment.
  • With transceivers and fiber basically any run length can be covered.
Cons of SFP+
  • Apparently, it is not a big deal for transmission within short distance.
  • For longer runs or runs that need to go through patch panels needs transceivers and fiber. Fiber itself is cheap but transceivers, termination, patch panels, and etc for fiber would cost a lot.
10GBASE-T Vs. SFP+

This passage would mainly demonstrate the difference between 10GbE base T and SFP+ options from the respective of technology, latency, and power consumption.

—Technology

Generally, 10GBase-T is cheaper and easier to deploy than the alternative SFP+ technologies. You can further compare these two different technologies in the following table:

Table1 Comparison between 10GBASE T and SFP Plus

—Latency

Low latency is paramount to ensure fast response time and reduce CPU idle cycles. That increases data center efficiency and ROI. With the increasing of using private cloud applications, the need for low latency is growing fast in large scale data centers.

When it comes to 10GBase-T, the PHY standard uses block encoding to transport data across the cable without errors. The standard specifies 2.6 microseconds for the transmit-receive pair, and the size of the block requires that latency to be less that 2 microseconds. SFP+ uses simplified electronics without encoding, and typical latency is around 300 nanoseconds (ns) per link. You can further compare them in the below table.

latency comparison

Basically, there are only slight differences between 10GBASE-T and SFP+ in terms of application latency. Relatively speaking, 10G SFP+ has lower latency than 10GBASE-T. High latency would exert negative influence on CPU and therefore limiting data center efficiency and increasing operational costs.

—Power Consumption

10GBase-T components today require anywhere from 2 to 5 watts per port at each end of the cable (depending on the distance of the cable) while SFP+ requires approximately 0.7 watt (regardless of distance). The difference is clearly shown in the below chart.

power consumption comparison

(Resource: http://www.datacenterknowledge.com)

Conclusion

Through this article, we are clear about the pros and cons of 10GBASE-T and SFP+ as well as their differences in technology, application latency and power consumption. It is evident that SFP+ is the right technology to ensure optimal performance with lowest latency and lower power usage in the data center. The cost saving becomes obvious when deploying from 1000 to 10,000 cables in the data center.

Cloud Core Switch—An Economic Choice for L3 Switch

MikroTik Switches have been popularly received favorable reviews, and this is inseparable with their keeping on the bleeding edge of switching technology. As a new member of MikroTik Smart Switch series, cloud core switch, also called cloud router switch, combing the best features of a fully functional router and a Layer 3 switch. That is to say, this cloud router switch works as both switch and router to connect the VLAN. This article would mainly discuss about cloud core switch, CRS226-24G-2S+RM switch and its connectivity solutions, as well as the reasons why they are economic choice for L3 switch.

About Cloud Core Switch

The cloud core switch, or cloud router switch, abbreviated as CRS, is a highly configurable switch, powered by RouterOS. It has 24 Gigabit Ethernet port. The Ethernet port 2-24 are switched, and the device can be accessed via these ports through the IP 192.168.88.1. Ethernet port 1 is configured as a DHCP client and has firewall on it. The SFP port is configured the same way as Ethernet 1, with a firewall and DHCP client on it. For the cloud router switch, there are nine models currently available. Here lists three different cases of the cloud core switch:

  • CRS125-24G-1S-2HnD-IN (integrated wireless, indoor case)
  • CRS125-24G-1S-IN (indoor case)
  • CRS125-24G-1S-RM (rackmount case)

MikroTik cloud router switch

Figure1: MikroTik cloud router switches(Resource: www.MikroTik.com)

Cloud Core Switch CRS226-24G-2S+RM

As one of the cloud core switches, CRS226-24G-2S+RM have been highly favored by most people. CRS226-24G-2S+RM is a fully functional layer 3 cloud router switch powered by Router OS, which is also available in 1U rackmount case. It comes with a special switch menu which includes all the specific configuration options for switches. It has 24 Gigabit ports and two SFP+ cages for 10G connectivity in which first SFP port supports 1.25G/10G modules and second port only 10G modules. Ports can be removed from the switch configuration and used for routing purposes if needed. The most distinctive feature of CRS226-24G-2S+RM is that uses a new class of switch chips, which allows us to have two SFP+ ports for 10G connectivity. The main features of this cloud core switch are listed in the following:

  • Fully manageable L3 switch, full wire speed switching
  • Configure ports as switch, or for routing
  • If required, full RouterOS power right there
  • SFP+ ports for 10G connectivity

CRS226-24G-2S+RM_big

Figure2: cloud core switch CRS226-24G-2S+RM(Resource: www.MikroTik.com)

Connectivity Solutions for CRS226-24G-2S+RM

As being mentioned, the cloud core switch CRS226-24G-2S+RM has 24 Gigabit ports and 2 SFP+ ports. For the twenty-four 10/100/1000 Ethernet ports, you could use both network cables and optical transceivers to connect. The transmission speed of Cat5 and Cat5e cables can be up to 100 Mb/s and 1G respectively. Besides, you can also use 10/100/1000BASE-T copper transceiver to make network connectivity. But it costs more than the network cables. In terms of 10G SFP+ ports, there are also two connectivity approaches. You can use both 10G SFP+ modules and 10G SFP+ DAC copper cable to connect. Relatively speaking, the 10G DAC cable is cheaper a lot than the 10G transceiver. But if transmission quality is your pursuit, and then 10GBASE SFP+ transceivers would be a good choice.

Why Are Cloud Core Switches Economic Choice for L3 Switch?

According to the above description, cloud core switches are powered by Router OS. RouterOS lets you add upper layer functionality. The cloud core switch is very far below wire speed when doing layer 3 or above. In fact, the cloud core switch is more of a bare-bones layer 2 switch that has an embedded low-horsepower router. In short, the switch features are useful for making bridges that work at wire speed, but they’re limited to simple forwarding and vlan handling. The bridge feature lets you glue almost anything together, and gives lots of filtering/manipulation tools, but it cannot perform at wire speed because it uses the main CPU. Last but not least, the average prices of Mikro Tik cloud core switches are not more than $150, you can check them by this link.

Conclusion

Cloud core/router Switch is a managed switch that runs RouterOS and SwitchOS, which delivers a high performance as a Layer 3 switch. They allow to manage port-to-port forwarding, apply MAC filter, configure VLANs, mirror traffic, apply bandwidth limitation and even adjust some MAC and IP header fields. The economic L3 switch including several switch models covering wide range applications, like enterprise network and home network.

Three Media Options for 10GbE in Data Centers

With the added network infrastructure complexity, power demands, and cost considerations, 10 Gigabit Ethernet (GbE) comes to network administrators’ thinking point. While 1GbE connection is able to handle the bandwidth requirements of a single traffic type, 10GbE has been preferred as the ideal solution by customers to meet current and future input/output (I/O) demands. Delivering more bandwidth, 10GbE simplifies the network infrastructure at the same time by consolidating multiple gigabit ports into a single 10gigabit connection.

Generally speaking, there are three media options for 10GbE: 10GBASE-CX4, SFP+, and 10GBASE-T. Each option has its own virtual point and downside in terms of cost, power consumption and distance reach. This paper analyzes these three options respectively, helping you understanding the pros and cons of current 10GbE media options.

10GBASE-CX4

10GBASE-CX4 was the first 10G copper standard published by 802.3 (as 802.3ak-2004), an early favorite standard for 10GbE deployments. Using the XAUI 4-lane PCS (Clause 48) and copper cabling similar to that used by InfiniBand technology, 10GBASE-CX4 is able to reach 15 meters. Practically, this option is limited by its heavy weight and expensive cables. In addition, the size of the CX4 connector prohibited higher switch densities required for large scale deployment. Larger diameter cables are purchased in fixed lengths, causing problems in managing cable slack. What’s more, the space isn’t sufficient to handle the larger cables.

10GBASE SFP+

SFP+ fiber optic cables and SFP+ direct attach cables (DACs) are all better solution than CX4.

10GBASE SFP+ Fiber Optic Cables

10GBASE-SR, 10GBASE-LR, 10GBASE-LRM are all specified to work through fiber optic cables, such as JD094B (shown below). This HP 10GBASE-LR SFP+ transceivers takes fiber as its transmission medium with distance up to 10km. Really, great for latency and distance, but fibers are expensive. Although they offer low power consumption, the project of laying fiber networks in data centers is limited due to the cost of the electronics largely. The fiber electronics can be four to five times more expensive than their copper counterparts, meaning that ongoing active maintenance, typically based on original equipment purchase price, is also more expensive.

JD094B, HP 10GBASE-LR SFP+ transceiver

10GBASE SFP+ DAC

DAC can be classified in to direct attach copper cable and active optic cable (AOC). On the one hand, SFP+ DAC is a lower cost option alternative to fiber, with its distance reaching flexible in 1m (eg. SFP-10G-AOC1M), 2m, 3m, 5m, 7m and so on. On the other, SFP+ DAC is not backward-compatible with existing 1GbE switches. Besides, this solution requires the purchase of an adapter card and requires a new top of rack (ToR) switch topology. And the cables are much more expensive than structured copper channels, and cannot be field terminated. All these factors make SFP+ DAC less popular the 10GBASE-T which will be discussed soon.SFP-10G-AOC1M, for short reach

10GBASE-T

10GBASE-T, or IEEE 802.3an-2006, is a standard released in 2006 to provide 10Gbit/s connections over unshielded or shielded twisted pair cables with distances up to 100metres (330 ft). Due to additional encoding overhead, 10GBASE-T has a slightly higher latency in comparison to most other 10GBASE standards. What’s more, 10GBASE-T offers the most flexibility, the lowest cost media. And because of its backward-compatibility with 1000BASE-T, 10GBASE-T can be deployed based on existing 1GbE switch infrastructures that are cabled with CAT6 and CAT6A (or above) cabling, keeping costs down while offering an easy migration path from 1GbE to 10GbE.

Conclusion

The deployment of 10GbE infrastructure should be much easier, with these media options in mind, coupled with your own such project considerations as cost, power consumption and distance reach. Fiberstore, as a professional fiber optic product supplier, offers a broad selection of fiber and copper cables, including SFP-10G-AOC1M mentioned above. For more information about 10GbE media options, you can visit Fiberstore.