Optical Fiber Selection for Network Interconnection

The emergence of Data Centers, Storage Area Networks and other computing applications drives the needs for ultra-high speed data interconnections and structured cabling. The interconnect media choices include wireless technology, copper cable and optical fiber cable. Fiber cable offers the highest bandwidth and supports the highest data rates. There are single-mode and multimode fiber types. Different types of fiber connect with fiber optic transceivers resulting in different performances and costs. So it’s important for the network designers to understand the fiber types and select the right fiber and corresponding fiber optic transceivers for network interconnection.

Optical Fiber Types

There are three main types of optical fiber suitable for network interconnection use:
9/125μm Single-mode fiber
50/125μm multimode fiber
62.5/125μm multimode fiber

The above numbers respectively mean the diameter of the glass core where the light travels and outside glass cladding diameter which is almost the same to most fiber types. So the difference of each fiber type is caused by the core diameter. It has great impact on system performance and system cost when balanced against network application needs. Two primary affected factors are attenuation and bandwidth.

Factors Affected by the Fiber Core Diameter

Attenuation is the reduction of signal power, or loss, as light travels through an optical fiber. Fiber attenuation is measured in decibels per kilometer (dB/km). The higher the attenuation, the higher rate of signal loss of a given fiber length. Single-mode fibers generally operate at 1310 nm (for short range) while multimode fibers operate at 850 nm or 1300 nm. Attenuation is not usually considered to be the main limiting factor in short rang transmissions. But it can cause big differences in high speed network such as 100Gb/s.

Bandwidth means the carrying capacity of fiber. For single-mode fiber, the modal dispersion can be ignored since its small core diameter. Bandwidth behavior of multimode fibers is caused by multi-modal dispersion during the light traveling along different paths in the core of the fiber. It has an influence on the system performance and data rate handling. Multimode fiber uses a graded index profile to minimize modal dispersion. This design maximizes bandwidth while maintaining larger core diameters for simplified assembly, connectivity and low cost. So manufacturers start to develop higher-performance multimode fiber systems with higher bandwidth.

System Costs: Single-mode and Multimode Fibers

A fiber optic transceiver usually consists the optical light sources, typically LED–light emitting diode and optical receivers. Since the core diameter size and primary operating wavelengths of single-mode fiber and multimode fiber are different, the associated transceiver technology and connectivity will also be different. So is the system cost.

To utilize the single-mode fibers generally for long distance applications (multi-kilometer reach), transceivers with lasers such as SFPP-10GE-LR (an SFP+ 1310nm 10 km transceiver supporting single-mode fibers) that operate at longer wavelengths with smaller spot-size and narrower spectral width. But these kinds of transceivers need higher precision alignment and tighter connector tolerance to smaller core diameters. Thus, it causes higher costs for single-mode fiber interconnections. To lower the cost, manufacturers produce transceivers based on VCSEL (vertical cavity surface emitting laser), for example, 10G-SFPP-SR (an SFP+ 850nm 300m transceiver supporting multimode fibers), which are optimized for use with multimode fibers. Transceivers applying low cost VCSEL technology to develop for 50/125μm multimode fibers, take advantage of the larger core diameter to gain high coupling efficiency and wider geometrical tolerances. OM3 and OM4 multimode fibers offer high bandwidth to support data rates from 10Mb/s to 100Gb/s.

Conclusion

Optical fiber is an easily-installed medium that is immune to electromagnetic interface and is also more efficient in terms of power consumption. What’s more, fiber optic cable can save space and cost with higher cabling density and port density over copper cabling. For single-mode fiber and multimode fiber, each one has its advantages and disadvantages. Network designers should better select the right fiber type and related fiber optic transceivers according to specific situations for higher system performance. Of course, cost is another important factor to be considered.

Originally published at www.fiber-optic-equipment.com

To Upgrade from 10G to 40G or not?

Some enterprises have built 10G network links. But one can never imagine how fast the network traffic is growing. There are many factors influencing the needs for high bandwidth, such as the rising populations of network subscribers, business and personal competitiveness, and the use of more video. In that way, 10G network is not able to meet the increasing needs. Then it pushes optical backbones and metro core networks to move to 40G network. However, to move to the high capacity network, many difficulties must be encountered in optical fiber networks. The question is to upgrade to 40G or not?

Difficulties of Upgrading to 40G

 

Cost Increases

The first challenge some service providers face when they consider moving to 40G technology is the cost of equipment. That’s the same when considering upgrading from 2.5G to 10G. The price of a 40G link is expensive than that of 4*10G links (as shown in Figure1). Equipment is one of the price determining factors. For example, a QSFP transceiver (the average market price of a JNP-QSFP-40G-LX4, Juniper Compatible 40GBASE-LX4 QSFP+ Transceiver, is more than US$ 2000.00) is more expensive than 4 SFP+ transceivers (the average market price of per FTLX8571D3BCL, Finisar Compatible 10GBASE-SR/SW SFP+ Transceiver, is about US$ 100.00). And some equipment may not be suitable for 40G networks, which means brand-new expensive equipment are needed.

Figure1. 10G to 40G

Optical Signal to Noise Ratio (OSNR) Drops

A unit of information, called a symbol, transmitted at 10G takes 100 pico seconds (100ps), and that symbol transmitted at 40G takes 25ps. This means the receiver translating the light back into a symbol deals with only 25% of the light of a 10G symbol. It causes 6 dB OSNR to drop. The OSNR is a measure of the strength of the signal. So the drop of 6dB means the link length will be decreased by 75%.

Chromatic Dispersion (CD) Increases

When a signal travels through a fiber, CD causes the pulses constituting the signal to spread in time. If this spreading is not compensated, these pulses will overlap. It means the signal is unusable. Comparing to 10G, this effect is 16 times more obvious at 40G. This creates a serious roadblock for 40G system operating.

Polarization Mode Dispersion (PMD) Increases

PMD is another problem which may be the most difficult to solve. This occurs due to infinitesimal imperfections in the circularity of the core of a fiber, which may be caused by the material itself, manufacturing process, or stress in the field created by bending or twisting. PMD is more capricious than predictable, and is very dependent on the qualities of the fiber. PMD can be influenced by factors including cable age or vintage, temperature of cable, cable design and cable manufacturer, etc.

Methods to Overcome Those Difficulties

To achieve 40G, new modulation schemes and approaches are being developed. The following tells about methods to overcome those difficulties.

Save Cost

We can discuss it from two sides: to buy or not to buy new equipment. First, to buy new equipment. As mentioned above, some optical equipment like transceivers from Google searching result are quite expensive. Thanks to the fast developed technology, manufacturers like Fiberstore (FS.COM) produce many compatible brands of which the price is quite low. The price of FS.COM JNP-QSFP-40G-LX4 is US$ 400.00 and FTLX8571D3BCL is US$ 16.00. Compared with the average market price, it’s obvious that FS.COM products have advantages on price. Now you may doubt that transceivers with such a low price will not have good quality. Then you don’t need to worry about it when you see FS.COM test assured program which provides world-class customer confidence in each fiber optics. Except transceiver optics module, FS.COM also has other cost-effective optical products for you to upgrade to 40G network.

Second, not to buy new equipment. Consider that many enterprises have spent a lot investments in 10G networks, many carriers would like to upgrade to 40G by just inserting new 40G cards into their existing 10G WDM systems to save cost. Here are three advisable strategies moving from 10G to 40G: Use the current fiber plant of the 10G routes, and not invest in expensive deployments of new fiber for 40G; Add 40G to currently installed equipment, so additional equipment is not required to sit alongside the 10G gear; Keep the network robustness and architecture intact, that is, the current line design, so that the numbers of OADM and ROADM nodes on any pathway in a ring or mesh do not need to be reduced.

Improve OSNR, CD, PMD

Problems related to OSNR, CD and PMD are caused by optics, the fiber quality and other optical equipment on one side. So you should better choose high quality equipment. On the other side, it depends on advanced optical technology bringing about simplified materials and better methods.

Conclusion

Although there are many challenges, the trend of upgrading from 10G to 40G is inevitable. With the above methods and more advanced fiber optical technologies, you should upgrade to 40G network without any hesitation. Besides, to help customers to achieve 40G smoothly, FS.COM researches and develops high-quality products with favorable price. And FS.COM will keep providing customers with more cost-effective equipment and better solutions to realize 40G upgrading.

Originally published at www.fiber-optic-equipment.com

What Should We Know About Cabling Before Deploying 10 Gigabit Ethernet?

Since the need for high data speed increases and the price of optical equipment becomes more affordable, many enterprises start to deploy 10 Gigabit Ethernet in their cooperate backbone, data centers to support high-bandwidth applications. But what should we know about cabling before deploying 10 Gigabit Ethernet?

Fiber Cabling Choices

Three factors should be considered for fiber cable deployment: fiber cable type, 10 Gigabit Ethernet physical interface and fiber optical transceiver module. The following tables show the standard fiber cables, physical interfaces, and transceiver modules applicable to 10 Gigabit Ethernet.

Fiber Cables	Multimode	OM1 fiber (62.5/125 μm)
		OM2 fiber (50/125 μm)
		OM3 fiber (50/125 μm)
	Single mode	9/125 μm fiber
Physical Interfaces	10GBase-LRM	Max distance 220m
	10GBase-S	Max distance 300m
	10GBase-L	Max distance 10km
	10GBase-E	Max distance 40km
	10GBase-Z	Max distance 80km
Transceiver modules	XENPACK	Large form factor
	X2	Smaller than XENPACK
	XFP	Smaller than X2
	SFP+	Smallest form factor

Note: the 10 Gigabit Ethernet physical interface type should be the same on both ends of the fiber link. For example, it is OK to deploy a fiber link with one XFP-10G-MM-SR optics on the left, and one SFP-10G-SR optics on the right. However, one SFP-10G-SR optics and one SFP-10G-LRM optics can’t connect together because of different physical types.

Copper Cabling

As switching standards copper cabling standards develop, copper cabling for 10GbE is more widely used. There are three different copper cabling technologies for 10 Gigabit Ethernet. Each one has different performances and prices.

First, 10GBase-CX4 is the first 10 Gigabit Ethernet copper standard. It’s relatively economical and allows for very low latency. But the form factor is too large for high density port counts in aggregation switches.

Second, Small Form-factor Plus (SFP+) is the latest standard for optical transceivers. 10 Gb SFP+Cu direct attach cables (DAC) directly connect into an SFP+ housing. It’s the best copper solution for servers and storage devices because it has low latency, small form factor and reasonable price.

Third, 10GBase-T is a fully IEEE compliant Ethernet transport technology standard, as defined by IEEE 802.3an-2006. 10GBase-T is to run 10 Gigabit Ethernet over CAT6a and CAT7 copper cabling up to 100 meters. 10GBase-T copper twisted-pair cabling can enable the earlier 10MB, 100MB and 1GB operation. However, 10GBase-T still needs to be improved on its price, power consumption and latency.

Media	Copper cable	Range (max)	Average Latency
CX4	Twinax	15m (49ft)	0.1 μs
SFP+ DAC	Twinax SFP+CU	10m (33ft)	0.1 μs
10GBase-T	CAT6 RJ45	30m(98ft)—50m (164ft)	>1.5 μs
	CAT6a RJ45	100m (98ft)	>1 μs
	CAT7 GG45	100m (98ft)	>1 μs

SFP+ Direct Attach Cables

SFP+ direct attach cables integrate SFP+ compatible connectors with a copper cable into a low-latency, energy-efficient, and low-cost solution. SFP+ direct attach cables offer the smallest 10 Gigabit form factor and a small cable diameter for higher density and optimized rack space in 10 Gigabit Ethernet (GbE) uplinks and 10 Gigabit Fiber Channel SAN and NAS input/output connections. To use SFP+ direct attach cables can save you a lot compared with fiber optic solutions. And it can still provide lower latency and save up to 50% power consumption per port than other copper twisted-pair cabling systems.

SFP+ direct attach cables can also provide enhanced scalability and flexibility. The cables connect several servers or storage devices together in a single rack. Thus, it reduces the use of intermediate patch panels. And it’s easy to move racks or deploy one rack at a time since the cabling outside of the rack is limited to the main switch connection.

Fiberstore offers comprehensive cabling solutions for 10 Gigabit Ethernet, including fiber cables, copper cables, and SFP+ direct attach cables and each one has various subcategories. Before deploying 10 Gigabit Ethernet, you need to consider factors of the performance, cost, power consumption and latency and choose the most suitable cabling solution.

Originally published at www.fiber-optic-equipment.com

How Does 10 Gigabit Ethernet Cabling Evolve?

The demand for high bandwidth promotes the development of data transmission technology. Ethernet standard continuously evolves to meet fast speed need, from 100BASE, 1000BASE to 10GBASE. Meanwhile, the data carrying technology also develops to provide great bandwidth for transporting data with low cost, such as the copper and fiber cables as well as optical transceiver modules.

Figure1. 10G Structured Cabling

Copper and Fiber

In 10 Gigabit Ethernet, copper and fiber are used to transport data. Each one has its own advantages and disadvantages.

Copper is more affordable and easy to install. It acts the best when used in short lengths, typically 100 meters or less. But when deployed over long distance, electromagnetic signal characteristics will influence its performance. Besides, bundling copper cabling can cause interference, which makes it difficult to employ as a comprehensive backbone. So copper cabling are widely used in PCs and LANs communication network instead of campus or long-distance transmission.

Compared with copper, fiber cabling is usually used for long distance communication among campus, and environments that need protection from interference, such as manufacturing areas. In addition, fiber cabling is more reliable and less susceptible to attenuation, which makes it suitable for data transmission distance over 100 meters. But fiber still has drawbacks. It’s more costly than copper.

The Evolution of Cabling Technologies

Since 10 GbE technologies have changed, so have the cabling technologies. There are two main standards: IEEE802.3ae and IEEE802.3ak. Factors covered in these standards like transmission distance and equipment being used are helpful to determine the cabling strategy.

• IEEE802.3ae

IEEE802.3ae standard updates the existing IEEE802.3 standard for 10GbE fiber transmission. The new standard defines several new media types for LAN, metropolitan area network (MAN) and wide area network (WAN) connectivity.

10GBASE-SR – it supports 10GbE transmission over standard multimode fiber (850 nm) for distances of 33 and 86 meters. The SR standard also supports up to 300 meters using the new 2000MHz/km multimode fiber (laser optimized). This one is the lowest-cost optics for 10GbE.

10GBASE-LR – it uses optics (1310nm) and supports single-mode fiber up to 10 km.

10GBASE-LX4 – it can support multimode fiber for distances up to 300 meters using Coarse Wavelength Division Multiplexing (CWDM). The LX4 standard also supports single-mode fiber for up to 10 Km. LX4 is more expensive than both SR and LR because it requires four times the optical and electrical circuitry in addition to optical multiplexers.

10GBASE-ER – it uses optics (1550nm) to support single-mode fiber up to 30 km.

• IEEE802.3ak / 10GBASE-T

10GBASE-T is the latest proposed 10GbE standard for use with unshielded twisted-pair (UTP) style cabling. This standard is to improve the performance and increase the transmission distance at a lower cost. Category 5 (Cat 5) and Category 6 (Cat 6) are the most common cabling systems being installed today. But Cat 5 can’t meet the bandwidth demands of 10GbE’s transmission. To meet the needs of 10GbE, manufacturers create Category 6A (Cat 6A), designed with existing Cat 6 cable but measured and specified to higher frequencies. In addition to Cat 6A, 10GBASE-T will operate on Category 7 (Cat 7) cables.

10GbE Transceivers

Except the cabling, transceivers also need to be considered for the network connectivity. Transceivers provide the interface between the equipment sending and receiving data. 10GbE has four defined transceiver types, including XENPAK, X2, XFP and SFP+ (Small Form-factor Pluggable Plus). These transceivers are pluggable and are compliant with 802.3ae standard.

Among them, SFP+ is the smallest 10G form factor. And it can interoperate with XENPAK, X2, XFP interface on the same link. Fiberstore provides a number of interfaces attempted to satisfy different objectives including support for MMF and SMF compatibility, such as SFP-10G-SR, SFP-10G-LR, SFP-10G-ER, SFP-10G-ZR, etc. For example, SFP-10G-SR transceiver module can support 300 meters data transmission distance over 850 nm multimode fiber. And SFP-10G-LR module supports the link length up to 10 kilometers over 1310 nm single mode fiber.

Figure2. SFP+ Transceiver

As the corresponding cabling technology gets great improvement, 10G network is becoming more affordable and pervasive. 10G network brings us higher speed. For 10G network connectivity, SFP+ transceivers are recommended to transport data over copper or fiber cabling.

Originally published at www.fiber-optic-equipment.com

Do You Know about QSFP LR4 PSM?

40GBASE QSFP+ (Quad Small Form-Factor Pluggable) transceivers are widely provided all round the world. These modules offer various high-density and low-power 40 Gigabit Ethernet connectivity options for data center, high performance computing networks, enterprise core and distribution layers and service provider applications. Among so many options, such as QSFP-40G-SR4, QSFP-40G-LR4, QSFP-40G-ER4, WSP-Q40GLR4L, etc., QSFP-40G-LR4 is well know as it can support link lengths of up to 10 km over single mode fiber. It enables high-bandwidth 40G optical links over 12-fiber parallel fiber terminated with MPO/MTP multifiber female connectors. The following will talk about one of QSFP-40G-LR4: QSFP-LR4-PSM (Parallel Single Mode).

Main Features

The QSFP LR4 PSM is a parallel 40G Quad Small Form-Factor Pluggable optical module. It offers increased port density and saves cost for total system. This module provides 4 independent transmit and receive channels. It can support 4x10 Gbps operation for an aggregate data rate of 40 Gbps on 10km of single mode fiber. It can also support 1x40G optical links over 12-fiber parallel fiber terminated with MPO/MTP connectors. But the ribbon cable with MPO/MTP connector should have proper alignment instead of being twisted. The module operates with single +3.3V power supply. With a 2-wire serial interface, it allows to send and receiver more complex control signals and to receiver digital diagnostic information. If there are some channels not used, then the unused channels can be shut down for maximum design flexibility. This transceiver module with form factor, optical/electrical connection and digital diagnostic interface is compliant with Multi-Source Agreement (MSA). It can meet the harshest external operating conditions including temperature, humidity and EMI interference (Electromagnetic Interface). For example, its operating case temperature ranges from 0 to 70 ℃。

Working Principle

This QSFP LR4 PSM is a parallel single mode optical transceiver with an MTP/MPO fiber ribbon connector. This transceiver module offers 4 transmit and 4 receive channels and each can support 10.3125 Gbps data rates (as shown in the following figure). The transmitter accepts electrical input signals compatible with Common Mode Logic (CML) levels. All input data signals are differential and internally terminated. The receiver converts the input signals via a photo detector array into parallel electrical output signals. And the outputs signals are also voltage compatible with Common Mode Logic (CML) levels. All signals are differential and support a data rates up to 10.3 Gbps per channel.

Each module offers 7 low speed hardware control pins (including the 2-wire serial interface): ModSelL (Module Select), SCL (Serial Clock), SDA (Serial Data), LPMode (Low Power Mode), IntL (Interrupt), ResetL, and ModPrsL (Module Present). Here just introduces several pins.

ModSelL is an input pin. When held low by the host, the module responds to 2-wire serial communication commands. The ModSelL allows the use of multiple QSFP+ modules on a single 2-wire interface bus–individual ModSelL lines for each QSFP+ module must be used.

SCL and SDA are required for the 2-wire serial bus communication interface and enable the host to access the QSFP+ memory map.

LPMode pin is used to set the maximum power consumption to protect the host. If the hosts are able to cool higher power modules, such modules will be accidentally inserted.

Fiberstore’s QSFP-40G-LR4 is designed for use in 40 Gigabit Ethernet links over single mode fiber. It’s compliant with QSFP+ MSA and IEEE 802.3ba 40GBASE-LR4. LR4 PSM is one of QSFP-40G-LR4 modules. All these cost-effective transceivers have to be tested before shipping to ensure full compatibility and to offer the best customer service. For more detailed information, please feel free to contact sales@fs.com.

Originally published at www.fiber-optic-equipment.com

Preparation for 40G/100G Migration

10G is now common in large enterprises. New network trends continue to drive the demand for high-speed Ethernet, such as the virtualization trend, network storage trend, I/O convergence trend, and data center network aggregation trend. So 40G and 100G as well as corresponding equipment are introduced into the market. The migration from 10G to 40G/100G is inevitable.

IEEE and TIA Standards

Before planning for migration to 40G/100G network, we should better know well about high-speed Ethernet. The following will talk about it from the side of standards. Because structured cabling systems design is always guided first by standards. The standards for 40G and 100G are significantly different from previous generations; active equipment and how information is transmitted are unique.

First, it’s IEEE standards. IEEE creates the standards that define performance parameters. IEEE 802.3ba 40Gb/s and 100Gb/s Ethernet is the only current standard that addresses the physical layer cabling and connector media maximums for 40/100G fiber channel requirements. IEEE 802.3ba-2010 standard was approved at the June 2010 IEEE Standards Board meeting. The standard is shown in the following table.

40G/100G Using MPO/MTP Interface
Second, it’s TIA (Telecommunications Infrastructure Standard). For data centers, TIA defines how to apply the parameters to structured cabling systems. It establishes design criteria including space and layout, cabling infrastructure, tiered reliability, and environmental considerations. The standard recommends using the highest capacity media available to maximize infrastructure lifespan.

1G and 10G networks use GBIC (Gigabit interface converter). For example, generally the transceiver SFP+ (small form-factor pluggable) is for 10G network. Later the fiber connectivity in high-speed active equipment becomes condensed and simplified. Transceivers for 40G and 100G are QSFP (quad small form-factor pluggable), CFP and CXP (100G form-factor pluggable). MPO/MTP is the designated interface for multimode 40/100G, and it’s backward compatible with legacy 1G/10G applications as well. Its small, high-density form factor is ideal with higher-speed Ethernet equipment.

Figure1. MPO/MTP Connector

40G and 100G Ethernet employ parallel optics. Data is transmitted and received simultaneously on MTP interfaces through 10G simplex transmission over each individual strand of the array cable.

After introducing some basics of the high-speed Ethernet, we’ll discuss the structured cabling system of migration to 40G and 100G networks in the simplest and most-effective way.

12- or 24-Fiber Cabling Infrastructure

The system includes configurations for 10G to 40G/100G networks over 12- or 24-fiber MTP cabling. What’s the difference between the two methods? Which one is better? The sections will compare the two from the sides of migration, density and congestion.

Migration To achieve the migration, components like trunks, harnesses, array cords, modules, and adapter plates are needed. With the 40G 12-fiber legacy configurations, a second trunk and another set of array harnesses will be needed to achieve 100% fiber utilization (as shown in Figure 2). For 100G, it also needs these additional components with 12-fiber legacy configuration. But with 24-fiber trunks, a single cable can support a 1G-100G channel and simplify network upgrades immensely (as shown in Figure 3). When equipment is upgraded, there is no need to install new trunks. In addition, limiting changes can reduce the inherent risks to network security and integrity.

Figure2. 12-Fiber Cabling

Figure3. 24-Fiber Cabling

Density The higher density connectivity, the more rack space for active equipment. Thus less floor space is needed. In this way, 24-fiber cabling has the obvious advantage. If the active equipment is configured for 24-fiber channel/lane assignments, there will be twice as as many connections with the same number of ports compared to 12-fiber.

Congestion The more connectivity you are able to run in the same footprint, the more crowded it can become at the rack or cabinet. Fewer trunks reduce cable congestion throughout the data centers. Using 24-fiber MTP trunks for the cable runs will save half the number of cables versus 12-fiber in the network. Runs carry a lighter load, fibers are easier to manage, and improved airflow reduces cooling costs. So 24-fiber MTP trunks offer a huge benefit.

Conclusion

The high-speed network will become more and more popular. It’s very important to know something about the migration to 40G/100G. To upgrade your network, 24-fiber MTP will be a better fiber cabling choice compared with 12-fiber. Do you prepare well for the great migration?

Originally published at www.fiber-optic-equipment.com

Which One Will You Choose for Your 40/100G Network, OM3 or OM4?

40G has been widely used in data centers. 100G will also come soon. To meet these high bandwidths, related fiber cables are needed. OM3 and OM4 can be used to transmit parallel optical signal. But what is their difference? Which one will you choose for your network?

Both OM3 and OM4 are laser optimized fiber. Their cores size is 50/125. Connectors are the same and both operate 850nm VCSELS (Vertical-Cavity Surface-Emitting Lasers) transceivers. So the difference lies in the construction of the fiber cable, which means OM4 cable has better attenuation and can operate at higher bandwidth than OM3.

Attenuation is the reduction in power of the light signal as it is transmitted (dB). Attenuation is caused by losses in light through the passive components, such as cables, cable splices, and connectors. As the connectors are the same, so the difference in OM3 and OM4 performance is in the loss (dB) in the cable. The maximum attenuation of OM3 allowed at 850 nm by the standards is less than 3.5 dB/km, and less than3.0 dB/km for OM4.

Another factor influencing the cable function is dispersion. Dispersion is the spreading of the signal in time due to the different paths the light can take down the fiber. It has two types: chromatic and modal. In multimode fiber transmission, chromatic dispersion is negligible and the modal dispersion is the limiting factor.

The modal dispersion determines the modal bandwidth that the fiber can operate at and this is the difference between OM3 and OM4. Modal bandwidth represents the capacity of a fiber to transmit a certain amount of information over a certain distance and is expressed in MHz*km. The higher the modal bandwidth the more information can be transmitted. The modal bandwidth of OM3 is 2700 megahertz*km while the mod0al bandwidth of OM4 is 4700 megahertz*km. Thus, OM4 allows the cable links to be longer.

Compared with OM3, OM4 has a lower attenuation and operates at a higher modal bandwidth. That means over OM4 less power is lost during the signal transmission and the signal can be transmitted further or through more connectors (which add to the losses). The following table shows the Ethernet distances at 850 nm supported by OM3 and OM4 respectively.

	1Gb	10Gb	40Gb	100Gb
OM3	1000m	300m	100m	100m
OM4	1000m	500m	150m	150m

So why is the standard for 40G only 100m on OM3 and 150m on OM4 compared to 300m and 500m for 10G? There are two reasons. First, when the IEEE 802 standard was created they decided to create a standard based on “relaxed” transceiver specifications so that smaller and lower cost transceivers could be used. Two functions of 10G transceivers (clock recovery and attendant re-timing) are absent in both QSFP+ (40G) and CFP (100G) devices. Second, the standard allows for transceivers with wider spectral width lasers which increase chromatic dispersion (pulse spreading). The quality of transceivers is also a factor.

Which will you choose for your 40/100G network, OM3 or OM4? Except the transmission distance and the cable costs, there are additional factors to consider such as the number of cross connects required and the mix of 40G port to 40G port and 40G port to 10G port. Because 40G signal is transmitted across eight pairs of fiber each with 10G. Similarly, it is important to take into account the likely location of future 100G equipment and the possible 100G to 100G, 100G to 40G and 100G to 10G connectivity requirement.

Originally published at www.fiber-optic-components.com