Archive | C-RAN RSS feed for this section

Meet the 5G Alternative: pCell

6 Feb
There’s a reason the US wireless operators just coughed up $45 billion on spectrum and that 5G is getting so much attention: Operators have a ceaseless need for more capacity in this age of smartphones, tablets and the Internet of Things. (See Hey Big Spenders! AT&T, Dish & VZ Splash Cash on Spectrum and Ericsson Testing 5G Use Cases, CFO Says.)

If you need further proof, look to Cisco Systems Inc. (Nasdaq: CSCO)’s venerable Visual Networking Index (VNI) released today, citing that mobile users across the globe cannot get enough of data, with 2.5 exabytes being consumed per month in 2014, a number Cisco expects to rise to 25 exabytes per month in 2019. An exabyte is one billion gigabytes or, in layman’s terms, a butt-load of data. (See Cisco’s Visual Networking Index and Cisco’s VNI Shines Light on Mobile Offload.)

I recently spoke with the CEO of an interesting startup that’s not waiting for 5G standards to be fleshed out, nor even hitching his technology to the 5G hype-wagon. He’s promising a solution to the spectrum crunch that is readily available today. The company is Artemis, and the technology is pCell, a centralized-radio access network (C-RAN) architecture Steve Perlman invented to use cell signal interference to bring high-power signals to individual mobile users.

The company isn’t new — it launched its product with a big PR splash last year, and it’s been working on the technology a decade longer than that. But Perlman says it’s finishing trials and testing now and gearing up for actual deployments. He attributes the lag time to getting over the credibility hump.

Indeed, the startup has had a tough time convincing operators that its technology works as advertised, bringing 25 times performance improvement from the same spectrum and the same devices they’re already using for LTE, without increasing costs substantially. He says that operators still can’t wrap their heads around it even when he shows them the technology working in front of their own eyes.

Analysts we spoke with shared the operators’ disbelief and added their own concerns about standards, scalability and working in the real world. The proof will be in the deployments that Perlman says are coming this year.

In the meantime, read up on pCell in our Prime Reading feature section hereon Light Reading to learn more about the technology, the promise and the challenges and to judge for yourself whether pCell is too good to be true or the magic bullet operators have been searching for. (See pCell Promises to Fix Spectrum Crunch Now.)



SK Telecom’s Network Evolution Strategies: Carrier aggregation, inter-cell coordination and C-RAN architecture

8 Oct

SK Telecom is the #1 mobile operator in Korea, with sales of KRW 16.6 trillion (USD 15.3 billion) in 2013, and with 50.1% of a mobile mobile subscription market share in 2Q 2014. It launched LTE service back in July 2011, and now more than half of its subscribers are LTE service subscribers, with 55.8% of LTE penetration as of 2Q 2014.
Due to LTE subscription growth, more advanced device features, and high-capacity contents, LTE networks are experiencing an unprecedented surge in traffic. To accommodate the flooded traffic, SK Telecom adopted LTE-A (Carrier Aggregation, CA) in 2013, and Wideband LTE-A (Wideband CA) in 2014 for improved network capacity.
As another effort to increase network capacity, the company made LTE/LTE-A macro cells a lot smaller, as small as hundreds of meters long, resulting in an increased number of cell sites. To save costs of building and operating the increased number of cell sites, it has built C-RAN (Advanced-Smart Cloud Access Network, A-SCAN, as called by SK Telecom) through BBU concentration since January 2012.
In 2014, SK Telecom began to introduce small cells (low-power small RRHs) in selected areas. As with macro cells, small RRHs have the same C-RAN architecture where they are connected to concentrated BBU pools through CPRI interfaces. SK Telecom calls it “Unified RAN (Cloud and Heterogeneous)”.
To prevent performance degradation at cell edges caused by introduction of small cells, SK Telecom developed HetNet architecture (known as SUPER Cell) where macro cells cooperate with small cells. The company, aiming to commercialize 5G networks in 2020, plans to commercialize SUPER Cell first in 2016, as a transitional phase to 5G networks.



Figure 1. SK Telecom’s Network Evolution Strategies
We analyzed SK Telecom’s network evolution strategies using the following three axes: 1) Carrier Aggregation (CA), 2) Inter-Cell Coordination, and 3) RAN Architecture in the Figure 1. Here, the CA axis shows how speeds have been and can be increased (n times) by expanding total frequency bandwidth aggregated. The Inter-Cell Coordination axis displays the company’s strategy to achieve higher speeds at cell edges by improving frequency efficiency. Finally, the RAN Architecture axis shows SK Telecom’s plan to switch to an architecture that would yield better LTE-A performance at reduced costs of building and operating RAN. Figure 2 is SK Telecom’s evolved LTE-A network, as illustrated according to the evolution strategies shown in Figure1.



Figure 2. SK Telecom’s LTE-A Evolution Network 



1. CA Evolution Strategies
CA is a technology that combines up to five frequencies in different bands to be used as one wideband frequency. It allows for expanded radio transmission bandwidth, which would naturally boost transmission speeds as much as the bandwidth is expanded. So, for example, if bandwidth is increased n times, then so is the transmission speed. Table 1 shows the LTE frequencies that SK Telecom has as of September 2014, totaling 40 MHz (DL only) across three frequency bands, which operate as Frequency Division Duplexing (FDD).
SK Telecom commercialized CA in June 2013 for the first time in the world, and then Wideband CA a year later in June 2014. 


It is now offering a maximum speed of 225 Mbps through the total 30 MHz bandwidth. As of May 2014, out of the total 15 million LTE subscribers, 3.5 million (23%) subscribers are using CA-enabled devices. Let’s see where SK Telecom’s CA is heading.


1.1 Combining More Bands: 3-band CA
3-band CA combines three frequency bands, instead of the current two, for wider-band transmission. Currently, SK Telecom has three LTE frequency bands, and is offering 2-band CA of 20 MHz or 30 MHz by combining two of the bands at once. This is because, although LTE-A standards technically support combining of up to five frequency bands, RF chips in  CA-enabled mobile devices available now can support combining of two bands only.  
3-band LTE devices are on the way and will be arriving in the market soon – sometime in early 2015 or by late 2014 at the latest. So, SK Telecom is planning to commercialize 3-band CA that combines all of its three frequency bands, just in time. The commercialization of 3-band CA is expected to increase transmission bandwidth to 40 MHz and data transmission rate to 300 Mbps. SK Telecom is also planning to combine three 20 MHz bands to further expand transmission bandwidth up to 60 MHz, and boost data transmission rate to 450 Mbps.


1.2 Femto Cell with CA
SK Telecom commercialized LTE Femto cell for the first time in the world in June 2012, to provide indoor users with more stable communication quality, and now is attempting to apply CA technology to Femto cell as well. The company completed a technical demonstration of LTE-A Femto cell in MWC 2014, proving it is capable to support 2-band CA. It will be conducting trial tests in a commercial network in late 2014 for final commercialization of the technology in 2015.


1.3 Combining Heterogeneous Networks: LTE-Wi-Fi CA
In July 2014, SK Telecom performed a technical demonstration of heterogeneous CA that combines LTE and Wi-Fi bands by using multipath TCP (MPTCP), an IETF standard. MPTCP is designed to combine more than one TCP flow (or MPTCP subflow) to make a single MPTCP connection, and send data through it. This technology is applied to a device and application server. In the demonstration, an MPTCP proxy server was used instead of an application server (Figure 3).    


Figure 3. LTE – Wi-Fi CA using Multipath TCP (MPTCP)
This technology will allow SK Telecom to combine i) its LTE bands that are currently featuring 2-band CA and ii) 802.11ac-based Giga Wi-Fi bands, together offering up to 1 Gbps or so. 
The detailed commercialization timeline is to be determined in accordance with the company’s plan for future development of MPTCP device and server.


1.4 Combining Heterogeneous LTE Technologies: FDD-TDD CA
This method enables operators to expand transmission bandwidth by combining two different types of LTE technologies: FDD-LTE and TDD-LTE. In a demonstration performed in Mobile Asia Expo in June 2014, SK Telecom successfully demonstrated FDD-TDD CA using ten 20 MHz bandwidths and 8×8 MIMO antenna showing 3.8 Gbps throughout. 


Quick Insight: Cloud RAN (C-RAN)

24 Oct


As all Telco engineers know that in a typical mobile deployment, each base station serves all the mobile devices within its reach. Each base station has its digital component manage its radio resources, handoff, data encryption and decryption and an RF component which transforms the digital information into analog RF. The RF elements are connected to a passive antenna that transmits the signals to the air. Each base station should be placed in the geographical center of its coverage area. But even when such locations are selected, the mobile operators may have difficulty in renting the real estate, finding proper powering options, securing the location and protecting the equipment from weather conditions. Those cell sites carry with them a continuous stream of OPEX to address the high rental rates for real estate, electrical expenses, cost of backhaul for the cell site and security measures to protect the location from intruders.

Enter the latest architectural paradigm : C RAN !!! The basic premise of Cloud RAN is to change the traditional RAN architecture so that it can take advantage of technologies like cloud computing, Software-Defined Network (SDN) approaches, and advanced remote antenna/radio head techniques.C-RAN architecture is not bound to a single RAN air interface technology. In essence, conventional terrestrial cell site base stations are replaced with remote clusters of centralized virtual base stations which can support up to a hundred remote radio / antenna units. This is achieved by centralizing RAN functionality into a shared resource pool or “cloud” (the digital unit – DU, or baseband unit – BBU) which is then connected via fibre to advanced remote radio heads (“Radio Units” – RU) sited in different geographical locations in order to provide full coverage of an area. The radical concept can even use banks of x86 servers to connect cellular calls rather than traditional wireless base stations.

From a business perspective,C-RAN will deliver significant reductions in Opex and Capex due to reduced upgrading costs. A major reason for this is the aggregation and pooling of the DU computing power which can be assigned specifically where needed e.g. the load situation over time and space for indoor/outdoor cells, am/pm hours, weekday/weekend, and so on. As a result, single cells do not need to be dimensioned for peak hour demands, but rather the processing power can be pooled and assigned on an on-demand basis. The processing power savings achieved should also leave processing headroom for any further potential technology enhancements (e.g., LTE-A features) without the need for further CAPEX. C-RAN skips the need for a high-bandwidth, low latency (X2), synchronized interface between the geographically distributed base station because the computing resources of the multiple transmission points’ BBUs are all located within the same hardware.

Furthermore, interference management will also benefit from C-RAN network architecture as technologies like dynamic eICIC schemes will be enabled, especially in a HetNet deployment.Heterogeneous networks will require small cells to be independent, intelligent and ubiquitous to avoid the cross- interference mayhem, yet be in synch and orchestrated with macro cells (including Cloud – RAN topology).Small cells are poised to become the most commonly used node for cellular access in the next-generation HetNet. C RANs will likely take their place beside traditional base stations and emerging small-cell base stations as another tool for building cellular nets.

According to Maravedis Cloud-RAN economics only be realized by harnessing standards to ensure interoperability and reduce cost. That, in turn, will create a whole new ecosystem, and operators must resist any attempts by their suppliers to hijack standards for software-defined networking or cell site equipment. Otherwise, this fledgling architecture will remain confined to a few pioneers with the resources to build their own ecosystems, like China Mobile.

China Mobile, the world’s largest carrier with 700 million subscribers, has been spearheading trials and plans to deploy systems as early as 2015. Japan’s NTT Docomo said it will follow in 2016, and a third unnamed carrier is now preparing plans for C-RANs. China Mobile aims to lower the cost of C-RANs to less than $30 per LTE sector, down from about $10,000 two years ago. It will start a second round of trials later this year using servers equipped with PCI Express cards to handle baseband processing. Each card will pack four FPGAs using silicon cores, each FPGA capable of handling 12 LTE sectors.

Pure C-RAN faces many barriers, such as over-reliance on fiber to link sites and basebands and immature standards, but most operators will inch towards C-RAN using hybrid models. Development of microwave fronthaul technologies will be critical to improve the C-RAN business model . Whatever the challenges C-RAN offers a revolutionary approach to next-generation cellular networks deployment, management and performance.

As MNOs face rising CAPEX bills to meet mobile data demand combined with falling ARPU, they must explore radical new network designs. With Cloud-RAN, they can virtualize baseband processing functions for hundreds of sites on a server or base station hotel. By consolidating individual Base-station processing into a single or regional server farm Investments on Cloud Radio Access Network (RAN) Infrastructure are expected to exceed $6 Billion by 2020, according to a new report from SNS Research. Distributed antenna technologies ( DAS ) will get a new lease on life, supporting coverage extension for C-RAN sites. This sector will open up $1.3bn in new revenues for antenna providers.

Sadiq Malik ( Telco Strategist )


Improving Capacity Coverage at the Network’s Edge

7 Aug

While all the commotion is in the small cell domain, let’s look at a traditional tail site and see what are the main ideas for capacity coverage improvement.

Just seven years ago, a tail site was connected with a single DS1 or E1. With LTE, we went up to 100-150 Mbps per site. Now we’re pitching 1Gbps per tail site. Mainly because LTE-Advanced can deliver 1Gbps using 100 MHz of spectrum with carrier aggregation.

Improvinf Capacity Coverage

The first question that arises in the minds of network engineers is, “Where am I supposed to get 100 MHz of spectrum?” The main option today is to re-farm old 2G and 3G spectrum to gain better spectral efficiency with  new gear.  The second option is to bid for new channels such as LTE3500. Upon release, 3.5GHz LTE will bring with it a potential 400 MHz for distribution in many countries.  In the meantime, the eco-system is not there yet but with 200MHz in FDD, 200 MHz in TDD – it is safe to say that it is likely to happen.

Higher spectrum is not very efficient for coverage but it is the right choice for small cells and tail sites.  We can reuse this spectrum many times due to its short range.  This spectrum was sub-optimal for WiMAX because it didn’t go very far but for small cells or a tail site that is covering a small area, 3.5GHz  is an excellent spectrum.

But this does not conclude all my capacity requirements in this particular location.  I am likely to have a RAN sharing model one way or another, with more than one mobile operator. This sharing concept requires a short explanation. The obvious trends are whole operations sharing such as EE in UK or a backhaul joint venture such as NetShare in Ireland. However the case of backhaul service provides (Carriers of Carriers – CoC or alternate access vendors – AAV) is very similar – it its about transparency and service differ nation. But it also means a need to serve additional spectrum slices per site in terms of capacity or marinating a more sophisticated timing scheme

And even more interesting, this is considered to be the best place to aggregate all my other small cells.  My offload, integrated, coordinated models, not to mention more sectors connected to a tail site –  aka, Distributed Base Station as we discussed in my past post.

All of these capacity requirements surface as this is my point of presence.  However, when we start talking about coordinated multi-point capabilities and carrier aggregation (CoMP). More coordination between the sites means higher capacities and lower latencies. Though it has yet to be seen how to implement these concepts in an ideal or non-ideal backhaul environment.  So in essence, it’s easy to see that 1 Gbps is going to be the new E1 for a tail site.

Stay tuned for part three of this conversation where I will discuss taking the distribution concept to the extreme with a move to Cloud-RAN (C-RAN).

In the meantime, for more information on how to increase capacity coverage, feel free to take a moment to view Ceragon’s new white paper on Capacity Coverage or feel free to reach out to me with any questions at


Cloud RAN Attracts Asian, European Carriers

17 Jul
SAN JOSE, Calif.— Three service providers are working on plans to deploy cloud radio access networks (C-RANs), a new approach to building cellular networks. The radical concept uses banks of x86 servers to connect cellular calls rather than traditional wireless base stations.

China Mobile, the world’s largest carrier with 700 million subscribers, has been spearheading trials and plans to deploy systems as early as 2015. Japan’s NTT Docomo said it will follow in 2016, and a third unnamed carrier is now preparing plans for C-RANs, said Gilad Garon, chief executive of Asocs Ltd. which sells silicon cores for modems.

“In last few months any doubts whether C-RAN would happen have gone away,” said Garon, whose company was chosen in February to supply baseband technology for China Mobile’s trials, in an interview. “Korea Telecom is involved, and we see interest in Europe primarily from Deutsche Telekom.”

China Mobile aims to lower the cost of C-RANs to less than $30 per LTE sector, down from about $10,000 two years ago. It will start a second round of trials later this year using servers equipped with PCI Express cards to handle baseband processing. Each card will pack four FPGAs using Asocs cores, each FPGA capable of handling 12 LTE sectors, said Garon.

The trial also will test Asocs’ MPL, a programming language for baseband chips. MPL lets developers call C-language libraries that will run jobs on ARM, Mips or x86 processors.

The next trial will use only Intel servers, but emerging low power ARM servers could be used in C-RANs in the future. “There’s still the cost of optics, software and servers, but the sheer DSP processing will be cheap,” said Garon.


Asocs' CR2100 cores will link to Intel Xeon chips in China Mobile's next C-RAN trial.

Asocs’ CR2100 cores will link to Intel Xeon chips in China Mobile’s next C-RAN trial.


C-RANs will likely take their place beside traditional base stations and emerging small-cell base stations as another tool for building cellular nets, said Gordon Mansfield, chairman of the Small Cell Forum and executive director of small cell solutions and radio access network delivery at AT&T Mobility.

“C-RANs’ biggest challenge is backhaul,” said Mansfield. “It requires extreme low latency and that requires fiber” which is expensive and not widely deployed in many parts of the world where traditional and small cells will be a better fit, he said.

With the exception of Intel, C-RANs also lack the support of the major chip makers, Mansfield, said. “Having general purpose hardware and software is the ultimate goal, and all the big, heavy-processing silicon providers will end up there I believe, but right now it’s too early to see who will come out as leaders,” he added.


%d bloggers like this: