Whatever 4G becomes, wireless operators around the world are preparing now.
Nothing ever stands still if you are a wireless operator. New technologies are always around the corner that demand attention, even ones that haven’t been fully defined.
Take 4G, for instance. No one has yet officially defined what a fourth generation wireless network will be, although a number of companies and organizations think they have an answer. But any use of the term 4G at this point is for marketing only.
Only the International Telecommunication Union (ITU) has the authority to put a “G” tag on a wireless access technology and the last time the ITU did that was in 2000. Nine years ago, it defined a 3G technology as being capable of peak data rates up to 2 mbps for stationary or walking users with at least 384 Kbps in a low-speed moving vehicle and 144 Kbps for a high-speed moving vehicle.
Technologies that fit the ITU’s 3G definition include cmda2000 1x and EV-DO, UMTS, TD-SCDMA, HSDPA and HSUPA. The TDD version of Mobile WiMAX also has been approved as a 3G technology, and the 3GPP’s Long Term Evolution (LTE) also was submitted for 3G inclusion. The latter, which both use OFDMA, may be designated 4G technologies when the ITU decides what that is.
Evolved Packet Core (EPC): A new all-IP mobile core for LTE |
Click Here or on Image to ENLARGE |
CO-EXISTENCE
But even though 4G hasn’t been fully defined yet, carriers around the globe are getting their networks in order for whatever it ultimately will be. One thing is certain – that whatever the designation, the technology won’t exist in a vacuum. Current operators will want the technologies to co-exist and even work together as the next generation is deployed.
“There’s a lot of life left in 3G,” points out analyst Andrew Seybold. Carriers won’t start rolling out LTE until late this year and the technology won’t be widespread for years and even then will need to be backwards compatible. Mark Price, senior telecom strategist for Openet, said operators looking to any 4G future “don’t want to put a stovepipe in place; they don’t want 4G to be isolated.”
Seybold said one way to insure that 4G isn’t a silo is for carriers using other access technologies to optimize their core network now so they’ll be ready for the next generation.
Using LTE as a 4G example, operators taking that path won’t have the radio access equipment to put into their networks until later this year, although most infrastructure vendors are in trials now. The first operators to deploy LTE, which includes Verizon Wireless, aim to launch the technology the end of 2009 or early 2010. But that doesn’t mean operators with LTE in their headlights are standing idly by. Many, if not most, are starting to upgrade their core networks and their backhaul to handle both the LTE technology and the increased traffic it is expected to bring.
FORGING AHEAD
“There’s quite a lot of activity around LTE now,” said Steve Vogelsang, vice president for mobility applications in Alcatel-Lucent’s IP Division. He said part of the planning and implementation process involves the packet core network. LTE also has a “flatter” architecture that reduces the number of controllers in a network, including the elimination of the radio network controller.
A flat infrastructure pushes much of the radio and routing functions to the edge of the network and connects them through an IP or packet core network. This reduces costly hardware and also makes such networks more efficient and simpler to manage.
Despite the changes, the LTE network and the legacy network must still talk to each other, whether the legacy network is based on CDMA2000 EV-DO as is the case with Verizon Wireless or W-CDMA/HSPA with AT&T Mobility.
|
The core network traditionally has been based on voice, which is circuit-switched going through a mobile switching center. When the GSM world added data to the networks with GPRS, these data networks “were glued onto the side,” according to Paul Steinberg, chief architect in Motorola’s Home & Networks Mobility unit. CDMA carriers didn’t have quite the same core network transition from 2G to 3G, but voice has remained circuit switched.
Voice may remain circuit switched for some time as legacy networks continue to be widely used, but the future with LTE and beyond is voice over IP. That will bring new challenges for the core network, especially for quality of service. Voice likely will be given higher priority over most data because users have a lower tolerance of diminished quality for voice calls.
The migration to 4G actually is taking place in the 3G core network, said Jon Morgan, director of product management at Starent. That’s because GSM operators had optimized their 2G core for data and are still evolving their packet core for the 3G technology W-CDMA/HSPA.
“We think they have to upgrade the packet core network to handle the HSPA part and then the migration story to LTE isn’t a hardware upgrade,’ Morgan said. “That’s what we’re hearing from our carrier customers. They want to do a software upgrade to go to LTE.”
The elements of the W-CDMA/HSPA network that need to be upgraded, according to Alcatel-Lucent’s Vogelsang, is that an evolved Node B (eNode B) would replace 3G’s base station and Node B. Then the Evolved Packet Core (EPC) would replace W-CDMA’s circuit-switched and packet-switched cores. The EPC includes four different functions—the serving gateway (SGW), packet data network gateway (PDN), mobility management entity (MME) and the policy charging rules function (PCRF).
Starent’s Morgan said carriers face a “double whammy” as their networks evolve because there will be huge increases in both throughput and in signaling. The signaling in an LTE network is handled by the MME, which replaces the functions of the serving gateway node in a 3G network. Starent combines the two in one platform to save signaling between the 3G and 4G networks, cutting costs and operational expense, he said.
3GPP Technologies Overview |
BSS/OSS CHANGES
One vital piece in the transition is billing for the various services running over the different network technologies. Openet’s Price said accessing an application server for content or gaming may have different characteristics depending on whether a 3G or 4G network is being used. Openet supplies software that allows operators to unify their charging policies across the technologies. “You don’t want to make massive changes to the old billing system,” Price said.
How billing will be handled in the new and old networks is one of the reasons most North American operators use Spirent and its network testing technology, according to James Brunson, director of converged network solutions. “We’re spending a lot of time looking at their BSS/OSS architecture and what changes need to be made,” he said.
Operators going to LTE or WiMAX are also using Spirent’s Landslide software to test the performance of the network elements, even during tests. Spirent also helps infrastructure vendors optimize their equipment in the labs and in the field.
The operators want to know if the new LTE eNode B they’re installing will be able to support tens of thousands of customers with the kind of throughput that LTE promises, he said.
REVOLUTION, EVOLUTION
You can start a good argument in LTE circles by debating the question whether or not the technology is “revolutionary” or “evolutionary.” The argument can get broken down between access and core network architecture. Spirent’s Brunson said most wireless carriers already have evolved at least part of their network to a packet core, although he argues that an all-IP (Internet Protocol) core is what is needed. Some operators have at least some all-IP elements, he said. The difference between a packet core and all-IP core are slight, but important. Motorola’s Steinberg said “packet core” is used to refer to mobile-specific technologies used in GPRS or EV-DO, with their respective packet switching technologies. An all-IP core refers to technology using switches and routers and servers that could be used in an Internet access model. GPRS, for example, uses a tunneling protocol to process packet data that isn’t native IP. Conversely, standardization work is under way to provide all-IP traffic to “tunnel” back to a legacy mobile network.Steinberg is one of those who thinks the transition from 3G to 4G is more of a revolution than an evolution because of the breadth and depth of changes being made to the network and the need to support legacy technologies.“It’s difficult to carry the old baggage forward,” he said. “It’s hard to have an evolution.”
Whether an evolution or revolution, it’s going to be costly. ABI Research has estimated the 18 carriers that have announced LTE deployment plans will spend $8.6 billion by 2013.
LTE Definitions |
Long Term Evolution (LTE) Evolved Packet Core (EPC) The Evolved Packet Core (EPC) highlights the growing importance of a common packet core across multiple access technologies. As many operators transition from disparate 3G specifications (UMTS and CDMA2000) to LTE, there is the potential for significant network simplification and cost savings, while also introducing greater efficiencies within the core network. The EPC defines a series of new network functions that flattens the architecture by reducing the number of nodes in the network, which promises to reduce capital and operational expenditures; thereby reducing the overall cost per megabyte of traffic running over the EPC, while improving network performance. System Architecture Evolution (SAE) Mobility Management Entity (MME) Serving Gateway (SGW) Packet Data Network Gateway (PGW) Gateway GPRS Support Node (GGSN) Serving GPRS Support Node (SGSN) Policy Charging Rules Function (PCRF) Evolved High Rate Packet Data (eHRPD) Sources: Starent Networks, 3GPP, 3G Americas, Wireless Week research |