Discussion of Offload Solutions
Mobile data offload emerging as 3G data core alternative
By Kevin Fitchard
January 29, 2010
Stoke joins Tellabs-WiChorus as latest company to offer Internet offload capabilities to rein in
rising mobile data core costs
Stoke made its mark last year as a base station and femtocell aggregator, but it turns out the
start-up 3G and 4G vendor has another trick up its sleeve. In April, Stoke will release its own
version of mobile data offload technology, designed to take Internet traffic bound for the Web off
of the mobile network before it hits the 3G core, potentially saving operators enormous costs in
core gateway investments.
Data offload isn’t a new concept with 3G networks. In fact, it’s at the heart of fixed mobile
convergence (FMC) technologies such as femtocells and dual-mode WiFi, which take voice and
data traffic off the carriers wide area radio access backhaul networks and shunts them on to
private broadband connections. FMC companies like Kineto Wireless have started specializing
in using WiFi as means of bypassing the 3G network, and gateway vendors like Genband have
built offload capabilities into the security gateways sitting at the edge of femtocell networks.
But those technologies look to bypass the access network and do little to address the growing
volumes of data traffic coming from the cell site itself. Furthermore, as that traffic increases,
carriers’ data revenues aren’t keeping pace, reducing their margins and forcing them to invest
ever greater amounts of money in infrastructure for lessening returns. On Thursday AT&T
(NYSE:T) reported that data traffic on its network had grown more than 200% in 2009, but its
data revenues for the fourth quarter had grown only 26%. Those trends are a growing concern
for mobile operators, and they’re looking for any way they can to maximize the amount of traffic
their networks can support for the smallest investment, said Barry Hill, vice president of sales
and marketing at Stoke.
“The mobile data traffic tsunami is hurting them in the radio access network, it’s hurting them
in backhaul, and it’s hurting them in the core,” Hill said. Carriers have little option in the
access and backhaul networks except to deploy more 3G channels, upgrade to more spectrally
efficient technologies and invest in bigger transport pipes. But the core network is a different
story: they can fundamentally change the way their networks treat data traffic, Hill said.
When 3G networks were first designed, the wireless industry was primarily a voice industry.
The standards were implemented with the expectation that customers would embrace data
services, but each customer would consume relatively small amounts. The opposite happened:
a minority of subscribers began consuming massive amounts of data. The serving General
Packet Radio Service (GPRS) support nodes (SGSNs) in the network, designed to scale for
subscribers, were suddenly being forced to scale for traffic volumes. And the GPRS gateway
serving nodes (GGSNs) that offloaded that traffic multiplied.
Most of that traffic, however, is bound for the public Internet, accessing content sitting on
someone else’s servers, while the carrier’s core network just passed those bits through. Mobile
data offload technologies, however, bypass the core network entirely. A box, in Stoke’s case its
SSX-3000 platform, sits in between the radio network controller and the SGSN, peering into the
packet stream to identify the ultimate destination of each data session. If a session is Internet
bound, and the operator has no value to add to the particular transaction, the offloader shunts
the traffic stream off of the carrier network onto the Internet, skipping the multiple waypoints it
would normally travel through in the carrier’s core.
Hill said a 3G operator with 10 million subscribers seeing a—rather conservative—350% growth
in data traffic over five years, would see its core investment balloon, starting with 24 SGSNs and
54 GGSNs in year one and ending with 144 SGSNs and 134 GGSNs In year five. Using mobile
data offload technologies, though, the same operator could run that network in year five with
just 12 SGSNs and 16 GGSNs, Hill said.
Stoke isn’t the only company that sees a big opportunity in data offload. Tellabs (NASDAQ:TLAB)
bought 3G-4G core vendor WiChorus last year, citing its mobile data offload as a key factor.
While Tellabs plans to get into the WiMax and HSPA-LTE core business, it also plans to
incorporate WiChorus technology into its backhaul portfolio as a means to bypass the core
entirely. But aside from start-ups and specialized vendors like Tellabs and Stoke, none of the
major vendors have been pressing forward with offload technologies.
The major vendors have been addressing the underlying problem. Direct tunneling
technologies such as Nokia Siemens Networks’ (NYSE:NOK, NYSE:SI) Internet-HSPA have
emerged and have been standardized. Direct tunneling solves the issue of scaling the SGSN as
it routes Internet traffic around it, but that Web-bound traffic is still traversing the core network
and the carrier must still deploy new GGSNs to handle the final hop to the public Internet.
Mobile data offload eliminates those steps entirely.
There are two reasons vendors haven’t moved beyond direct tunneling, Hill said: “One is a
technical reason and the other is completely and utterly commercial.” On the technical side,
3G networks aren’t IP networks—slicing your way through the layers of protocols to inspect the
control plane isn’t an easy thing to do.”You only get to see the IP packet when it’s on the other
side of the GGSN,” Hill said.
On the commercial side, it’s a simple matter of turf, Hill said. Vendors make a lot of money
selling core elements to operators. Why develop a technology that will eat enormously into your
SGSN and GGSN sales? You can’t blame vendors for dragging their feet though. Operators are
just as culpable. After all the stratified 3G architecture is the one they asked for. Operators have
also harbored the dream of the content provider of record for their customers, so the 3G network
was designed to accommodate a business model in which an operator provided most services
and added their little bit of a value to services provided by third parties. Offloading traffic before
it ever sees an SGSN would be embracing the dumb pipe model.
But Hill said he thinks operators have had a change in heart in recent years as mobile data has
exploded. While they’re not necessarily embracing the bit pipe model, operators have come to
the realization that a good portion of their traffic will be services they can add no value to. And
if they can’t add any value—or collect any incremental revenues—why pay for the
infrastructure to carry it?
See: http://www.item.ntnu.no/fag/tm8100/Pensumstoff2004/GPRS_Tutorial.pdf
Next
Mobile data offload emerging as 3G data core alternative
By Kevin Fitchard
January 29, 2010
Stoke joins Tellabs-WiChorus as latest company to offer Internet offload capabilities to rein in
rising mobile data core costs
Stoke made its mark last year as a base station and femtocell aggregator, but it turns out the
start-up 3G and 4G vendor has another trick up its sleeve. In April, Stoke will release its own
version of mobile data offload technology, designed to take Internet traffic bound for the Web off
of the mobile network before it hits the 3G core, potentially saving operators enormous costs in
core gateway investments.
Data offload isn’t a new concept with 3G networks. In fact, it’s at the heart of fixed mobile
convergence (FMC) technologies such as femtocells and dual-mode WiFi, which take voice and
data traffic off the carriers wide area radio access backhaul networks and shunts them on to
private broadband connections. FMC companies like Kineto Wireless have started specializing
in using WiFi as means of bypassing the 3G network, and gateway vendors like Genband have
built offload capabilities into the security gateways sitting at the edge of femtocell networks.
But those technologies look to bypass the access network and do little to address the growing
volumes of data traffic coming from the cell site itself. Furthermore, as that traffic increases,
carriers’ data revenues aren’t keeping pace, reducing their margins and forcing them to invest
ever greater amounts of money in infrastructure for lessening returns. On Thursday AT&T
(NYSE:T) reported that data traffic on its network had grown more than 200% in 2009, but its
data revenues for the fourth quarter had grown only 26%. Those trends are a growing concern
for mobile operators, and they’re looking for any way they can to maximize the amount of traffic
their networks can support for the smallest investment, said Barry Hill, vice president of sales
and marketing at Stoke.
“The mobile data traffic tsunami is hurting them in the radio access network, it’s hurting them
in backhaul, and it’s hurting them in the core,” Hill said. Carriers have little option in the
access and backhaul networks except to deploy more 3G channels, upgrade to more spectrally
efficient technologies and invest in bigger transport pipes. But the core network is a different
story: they can fundamentally change the way their networks treat data traffic, Hill said.
When 3G networks were first designed, the wireless industry was primarily a voice industry.
The standards were implemented with the expectation that customers would embrace data
services, but each customer would consume relatively small amounts. The opposite happened:
a minority of subscribers began consuming massive amounts of data. The serving General
Packet Radio Service (GPRS) support nodes (SGSNs) in the network, designed to scale for
subscribers, were suddenly being forced to scale for traffic volumes. And the GPRS gateway
serving nodes (GGSNs) that offloaded that traffic multiplied.
Most of that traffic, however, is bound for the public Internet, accessing content sitting on
someone else’s servers, while the carrier’s core network just passed those bits through. Mobile
data offload technologies, however, bypass the core network entirely. A box, in Stoke’s case its
SSX-3000 platform, sits in between the radio network controller and the SGSN, peering into the
packet stream to identify the ultimate destination of each data session. If a session is Internet
bound, and the operator has no value to add to the particular transaction, the offloader shunts
the traffic stream off of the carrier network onto the Internet, skipping the multiple waypoints it
would normally travel through in the carrier’s core.
Hill said a 3G operator with 10 million subscribers seeing a—rather conservative—350% growth
in data traffic over five years, would see its core investment balloon, starting with 24 SGSNs and
54 GGSNs in year one and ending with 144 SGSNs and 134 GGSNs In year five. Using mobile
data offload technologies, though, the same operator could run that network in year five with
just 12 SGSNs and 16 GGSNs, Hill said.
Stoke isn’t the only company that sees a big opportunity in data offload. Tellabs (NASDAQ:TLAB)
bought 3G-4G core vendor WiChorus last year, citing its mobile data offload as a key factor.
While Tellabs plans to get into the WiMax and HSPA-LTE core business, it also plans to
incorporate WiChorus technology into its backhaul portfolio as a means to bypass the core
entirely. But aside from start-ups and specialized vendors like Tellabs and Stoke, none of the
major vendors have been pressing forward with offload technologies.
The major vendors have been addressing the underlying problem. Direct tunneling
technologies such as Nokia Siemens Networks’ (NYSE:NOK, NYSE:SI) Internet-HSPA have
emerged and have been standardized. Direct tunneling solves the issue of scaling the SGSN as
it routes Internet traffic around it, but that Web-bound traffic is still traversing the core network
and the carrier must still deploy new GGSNs to handle the final hop to the public Internet.
Mobile data offload eliminates those steps entirely.
There are two reasons vendors haven’t moved beyond direct tunneling, Hill said: “One is a
technical reason and the other is completely and utterly commercial.” On the technical side,
3G networks aren’t IP networks—slicing your way through the layers of protocols to inspect the
control plane isn’t an easy thing to do.”You only get to see the IP packet when it’s on the other
side of the GGSN,” Hill said.
On the commercial side, it’s a simple matter of turf, Hill said. Vendors make a lot of money
selling core elements to operators. Why develop a technology that will eat enormously into your
SGSN and GGSN sales? You can’t blame vendors for dragging their feet though. Operators are
just as culpable. After all the stratified 3G architecture is the one they asked for. Operators have
also harbored the dream of the content provider of record for their customers, so the 3G network
was designed to accommodate a business model in which an operator provided most services
and added their little bit of a value to services provided by third parties. Offloading traffic before
it ever sees an SGSN would be embracing the dumb pipe model.
But Hill said he thinks operators have had a change in heart in recent years as mobile data has
exploded. While they’re not necessarily embracing the bit pipe model, operators have come to
the realization that a good portion of their traffic will be services they can add no value to. And
if they can’t add any value—or collect any incremental revenues—why pay for the
infrastructure to carry it?
See: http://www.item.ntnu.no/fag/tm8100/Pensumstoff2004/GPRS_Tutorial.pdf
Next
 | |  | |  | |  |