Is Xenpak Past It?

Xenpak, a popular form factor for 10-Gbit/s optical modules, is starting to look a little old and flabby.
Two newer, leaner form factors are going to steal Xenpak's market share in short-reach applications -- the largest market by volume -- according to a commentary from RHK Inc. published last week.
This shouldn't really come as a surprise. Xenpak, which has been around for more than a year now, has previously been criticized for being too cumbersome (see Sizing Up Xenpak ). The module was originally designed to dissipate the heat generated by long-reach lasers. But short-reach optics operate at lower powers, so they don't require all that cooling. In any case, laser technology has improved since Xenpak first came on the scene.
"We shouldn't penalize lower power, shorter reach transceivers by sticking them in big, bulky packages designed for stuff that drives a signal out 50 miles," contends Drew Lanza, an industry veteran of 15 years, now with Morgenthaler, a VC firm.
That's not to say that Xenpak wasn't a significant advance when it was first introduced (see Agilent, Agere Drive 10-Gig Ethernet ). It was one of the first modules to be pluggable into the side of a board, rather than from the top, allowing modules to be changed without taking cards out of the chassis. But technology moves quickly.
Competing multisource agreements (MSAs) to Xenpak were announced in March within a few days of each other. First was XFP, which is supported by a 44-vendor group including Broadcom Corp. (Nasdaq: BRCM), Brocade Communications Systems Inc. (Nasdaq: BRCD), Finisar Corp. (Nasdaq: FNSR), and JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU) (see 10 Vendors for 10 Gig). The XFP group announced the addition of 33 new members today (see XFP Touts Progress).
The other contender is XPAK, which was introduced by Intel Corp. (Nasdaq: INTC), Infineon Technologies AG (NYSE/Frankfurt: IFX), and Picolight Inc. (see Trio Announce 10-Gig MSA).
XPAK is a more direct "upgrade" for Xenpak because it sports the same XAUI (pronounced "zowie") electrical interface, and 70-pin connector. At the recent OFC conference, Picolight had a mockup showing that it is possible to cram 20 XPAK modules into the same space as eight Xenpak modules. That's achieved by making the XPAK modules fractionally less wide than Xenpak and then stacking them in pairs above and below the line card.
"XPAK's strength lies in its similarities to Xenpak as a classic second-generation product," says RHK program director Karen Liu. Systems vendors that have already designed in Xenpak, should be able to change to XPAK with minimum changes to their board designs.
XFP is a different fish, being a serial 10-Gbit/s solution that requires a physical-layer chip (Ethernet MAC or Sonet framer) to be placed on the line card (instead of being inside the module as with Xenpak and XPAK).
Placing the chip outside the module has a couple of potential advantages. One, there's less to put into the module, so it can be really small -- XFP is the same size as the popular SFP (small form-factor pluggable) connector used for transmitters. And two, it's protocol independent, which equates to a larger potential market and the promise of being able to manufacture it cheaply in volume.
Adoption of XFP is further out, reckons Liu. It will be a while before chip makers deliver the supporting electronic chips that make it work, she says. In addition, routing 10-Gbit/s signals on a printed circuit board is not easy -- keeping electromagnetic interference within tolerable levels will be a challenge for the systems houses. Indeed many of the companies supporting XFP are chip makers, which are involved in order to help solve exactly these issues.
In addition, the success of an MSA depends on its supporters. Reportedly, the main reason that Xenpak drew such widespread support was because Cisco Systems Inc. (Nasdaq: CSCO) was demanding the new high-speed modules. Agere Systems (NYSE: AGR) and Agilent Technologies Inc. (NYSE: A) stepped up to bring Xenpak to market, abandoning another 10-gig module effort, the XGP group (now dormant) in order to do so.
Both XPAK and XFP have signed up some heavy hitters. XPAK has the support of storage vendors McData Corp. (Nasdaq: MCDTA) and QLogic Corp. (Nasdaq: QLGC), as well as Intel, which is both a manufacturer and potential customer for the modules. As noted, XFP is supported by Brocade, and EMC Corp. (NYSE: EMC)also storage vendors.
— Pauline Rigby, Senior Editor, Light Reading
http://www.lightreading.com
Two newer, leaner form factors are going to steal Xenpak's market share in short-reach applications -- the largest market by volume -- according to a commentary from RHK Inc. published last week.
This shouldn't really come as a surprise. Xenpak, which has been around for more than a year now, has previously been criticized for being too cumbersome (see Sizing Up Xenpak ). The module was originally designed to dissipate the heat generated by long-reach lasers. But short-reach optics operate at lower powers, so they don't require all that cooling. In any case, laser technology has improved since Xenpak first came on the scene.
"We shouldn't penalize lower power, shorter reach transceivers by sticking them in big, bulky packages designed for stuff that drives a signal out 50 miles," contends Drew Lanza, an industry veteran of 15 years, now with Morgenthaler, a VC firm.
That's not to say that Xenpak wasn't a significant advance when it was first introduced (see Agilent, Agere Drive 10-Gig Ethernet ). It was one of the first modules to be pluggable into the side of a board, rather than from the top, allowing modules to be changed without taking cards out of the chassis. But technology moves quickly.
Competing multisource agreements (MSAs) to Xenpak were announced in March within a few days of each other. First was XFP, which is supported by a 44-vendor group including Broadcom Corp. (Nasdaq: BRCM), Brocade Communications Systems Inc. (Nasdaq: BRCD), Finisar Corp. (Nasdaq: FNSR), and JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU) (see 10 Vendors for 10 Gig). The XFP group announced the addition of 33 new members today (see XFP Touts Progress).
The other contender is XPAK, which was introduced by Intel Corp. (Nasdaq: INTC), Infineon Technologies AG (NYSE/Frankfurt: IFX), and Picolight Inc. (see Trio Announce 10-Gig MSA).
XPAK is a more direct "upgrade" for Xenpak because it sports the same XAUI (pronounced "zowie") electrical interface, and 70-pin connector. At the recent OFC conference, Picolight had a mockup showing that it is possible to cram 20 XPAK modules into the same space as eight Xenpak modules. That's achieved by making the XPAK modules fractionally less wide than Xenpak and then stacking them in pairs above and below the line card.
"XPAK's strength lies in its similarities to Xenpak as a classic second-generation product," says RHK program director Karen Liu. Systems vendors that have already designed in Xenpak, should be able to change to XPAK with minimum changes to their board designs.
XFP is a different fish, being a serial 10-Gbit/s solution that requires a physical-layer chip (Ethernet MAC or Sonet framer) to be placed on the line card (instead of being inside the module as with Xenpak and XPAK).
Placing the chip outside the module has a couple of potential advantages. One, there's less to put into the module, so it can be really small -- XFP is the same size as the popular SFP (small form-factor pluggable) connector used for transmitters. And two, it's protocol independent, which equates to a larger potential market and the promise of being able to manufacture it cheaply in volume.
Adoption of XFP is further out, reckons Liu. It will be a while before chip makers deliver the supporting electronic chips that make it work, she says. In addition, routing 10-Gbit/s signals on a printed circuit board is not easy -- keeping electromagnetic interference within tolerable levels will be a challenge for the systems houses. Indeed many of the companies supporting XFP are chip makers, which are involved in order to help solve exactly these issues.
In addition, the success of an MSA depends on its supporters. Reportedly, the main reason that Xenpak drew such widespread support was because Cisco Systems Inc. (Nasdaq: CSCO) was demanding the new high-speed modules. Agere Systems (NYSE: AGR) and Agilent Technologies Inc. (NYSE: A) stepped up to bring Xenpak to market, abandoning another 10-gig module effort, the XGP group (now dormant) in order to do so.
Both XPAK and XFP have signed up some heavy hitters. XPAK has the support of storage vendors McData Corp. (Nasdaq: MCDTA) and QLogic Corp. (Nasdaq: QLGC), as well as Intel, which is both a manufacturer and potential customer for the modules. As noted, XFP is supported by Brocade, and EMC Corp. (NYSE: EMC)also storage vendors.
— Pauline Rigby, Senior Editor, Light Reading
http://www.lightreading.com
EDUCATIONAL RESOURCES


FEATURED VIDEO
UPCOMING LIVE EVENTS
June 6-8, 2023, Digital Symposium
June 21, 2023, Digital Symposium
December 6-7, 2023, New York City
UPCOMING WEBINARS
May 30, 2023
Maximizing the Potential of 5G with Active Assurance
June 14, 2023
How do We Capture the 6G Experience?
June 14, 2023
The Power of Wholesale Order Automation: How New Advancements in Intercarrier Commerce Can Transform Your Business.
June 20, 2023
5G standalone for breakout growth and efficiency
June 21, 2023
Cable Next-Gen Europe Digital Symposium
June 22, 2023
Next-Gen PON Digital Symposium
Webinar Archive
PARTNER PERSPECTIVES - content from our sponsors
Is The Traditional PayTV Provider Being Squeezed Out?
By Terry Doyle for Enghouse Networks
Huawei Unveils the F5G Premium All-Optical Network, Enabling ISPs to Stride into the High-Growth Era
By Huawei
All Partner Perspectives