Tag Archives: QoE

LTE Femto Gateway with X2 Broker

14 Mar
 An LTE femtocell* (HeNB) is an ultra-small cellular base station that connects to a mobile operator’s LTE core network via broadband Internet. Using this femtocell, a mobile operator can eliminate indoor shadowing areas, thereby extending LTE service coverage and improving call quality.

* Femtocell and HeNB are interchangeable, and so are Femtocell Gateway and HeNB Gateway in this document.

The mobile operator can benefit from the femtocell as it allows LTE traffic to be distributed between macro eNB and femtocells at home and also at indoor and outdoor hotspots in crowded places like coffee shops, restaurants, bus stop, malls, schools, and so on. This helps the operator to effectively reduce loads at macro cells and in the backhaul, and provide its users with better QoE.

The beauty of LTE femtocell is that, as all it takes is simply connecting existing broadband Internet to an ultra-small base station, it gives the advantage of quick deployment. It also minimizes additional costs and burdens that may be imposed in case of building macro cells, in relation to installation site acquisition, site rental, power supply, construction of backhaul network, etc. Such benefits make it one of the most cost-effective ways to expand coverage and capacity in an LTE network.

Figure 1-1. Key values provided by femtocell in 4G era

Years ago, mobile operators started building macro LTE networks, and have always been in the quest for solutions to shadowing areas and high costs of operating multiple networks (2G, 3G and 4G) since then. Recently, operators are pursuing a strategy to i) provide uninterrupted voice coverage without relying on legacy networks like 2G or 3G by introducing small cells in shadowing areas and supporting seamless handover between them and macro cells, and ii) ultimately migrate into an all LTE network through gradual replacement of legacy networks.

That is, many operators are pushing forward with this strategy to minimize the total OPEX of the entire network by operating only one LTE network instead of multiple mobile networks. Femtocells are considered the most likely candidate to serve this purpose.

Meanwhile, operators without 2G or 3G, but with LTE macro network, are also active in introducing LTE femtocells in their networks as a cost-effective solution to enhance LTE coverage and capacity.

Here, what concerns the operators most is “uncertainty that can be caused while these femtocells (HeNB). Unlike existing macro cells, if deployed in a large scale – in tens or hundreds of thousands, these cells can cause unpredictable, operational risk while interworking with legacy LTE systems (EPC, eNB, etc.).”

SMEC’s Femto GW (HeNB-GW), designed to work as a sponge to absorb such uncertainty and risk, helps to operate the femto network just as stably as macro networks.

Chapter 2 will look into the benefits and issues of HeNB-GW, and chapter 3 will introduce HeNB-GW solution of SMEC, specifically X2 broker feature in details. Chapter 4 will summarize the benefits of the SMEC solution.


2. HeNB-GW: Benefit and Issues

2.1 Benefits

Table 2-1 summarizes issues in connecting HeNBs directly to MME, without HeNB-GW, as compared to the benefits of deploying HeNB-GW.

Table 2-1. Benefits of deploying HeNB-GW

2.2 Issues – X2 Handover Support

Mobile operators prefer X2 handover that uses just X2 interface between eNBs to more complicated S1 handover that increases loads at MME. As seen in Figure 2-1(a), the more HeNBs are deployed, the more hand-in and hand-out activities are performed between macro eNBs and HeNBs (particularly outdoors).

This means even more loads are caused at MME and S-GW by S1 handover, affecting the reliability of the LTE network. For more reliable, secured operation of the LTE core network, X2 handover without MME’s intervention is essential in a femto network (Figure 2-1(b)).

Figure 2-1. Handover options between macro eNB and HeNBs: S1 vs. X2


Figure 2-2. Issues: scalability and uncertainty

But in reality, supporting X2 handover in a femtocell environment is not easy because of possible scalability and instability issues. If existing macro eNBs establish X2 connections directly with a large number of HeNBs, scalability can be compromised due to the limit in the number of X2 connections that can be managed (Figure 2-2(a)).

For X2 handover, existing MME and eNB must interact directly with HeNBs (S1-MME, X2), and this process can bring about instability between the two (Figure 2-2(b)). Also, configuring X2 GW requires upgrade of eNBs and HeNBs all to R-12, consequently aggravating the complexity of the network even further.

These issues have been an obstacle standing in the way of applying X2 handover between macro eNB and HeNB in the commercial network. The HeNB-GW solution by SMEC is designed to address these issues. We will learn how in chapter 3.


3. SMEC HeNB-GW Solution


The Figure 3-1 describes a high level view of LTE network with femtocell and SMEC HeNB-GW. SMEC HeNB-GW can provide:

  • Virtual eNB (eNB ID based HeNB grouping)
  • X2 service broker (X2 proxy between eNB and HeNB)
  • S1 and X2 handover between eNB and HeNB
  • S1 signaling and bearer aggregation with SeGW functionality

Figure 3-1. SMEC HeNB-GW architecture

SMEC HeNB-GW, technologically based on virtual eNB concept, can group a number of HeNBs for management by group. Each virtual eNB, capable of aggregating 256 HeNBs, functions as a logical HeNB GW, providing S1 interface to EPC and HeNBs, and X2 interface to macro eNB and HeNB. From a S1 interface point of view, MME and S-GW see virtual eNB as ‘one macro eNB’, and HeNB sees it as ‘MME and S-GW’. Virtual eNB provides the following functionalities in respect of S1 interfaces:

  • Relaying UE-associated S1AP messages between MME and HeNB
  • Terminating non-UE associated S1AP procedures towards HeNB and towards MME
  • Terminating S1-U interfaces with HeNB and with S-GW

Virtual eNB, as a logical macro eNB, provides X2 interfaces. From X2 interface point of view, the macro eNB sees virtual eNB as an eNB with 256 cells that offers following functionalities:

  • Providing X2 interfaces between macro eNB and HeNBs
  • Terminating non-UE associated X2AP messages between eNB and HeNB
  • Converting UE-X2AP-ID between eNB and HeNB
  • Routing UE-associated X2AP messages between eNB and HeNB

3.2 X2 Service Broker

SMEC HeNB-GW features X2 service broker for complexity and stability issues as seen in Figure 2-2. As shown in Figure 3-2(b), each HeNB establishes X2 connection with virtual eNB (acting as a ‘X2 service broker’) at SMEC HeNB-GW, and macro eNB establishes only one X2 connection with the virtual eNB.

This X2 aggregation function provided by X2 broker drastically reduces the number of X2 connections needed between macro eNB and  HeNBs (256 X2 connections to only one X2 connection). SMEC HeNB-GW makes existing macro eNBs recognize it as another regular macro eNB, by hiding all the HeNBs behind its back.

Existing MME and eNB must interact directly with HeNBs (S1-MME, X2) for X2 handover, etc., and this can bring about instability between the two. X2 broker, upon receiving S1 and X2 messages from HeNB, modifies  the messages as if it is eNB itself, and sends them to MME and eNB. This ensures the stability of the LTE core network and eNB remains unaffected.

As a result, network complexity and unstability anticipated by deployment of HeNB can be significantly decreased, and kept as low as in existing macro eNB network. LG U+, a South Korean LTE network operator, has already deployed SMEC’s HeNB-GW, applying X2 handover between macro eNB and HeNBs in its commercial network. The company has been able to keep the load level at MME at a minimum and provide uninterrupted VoLTE service across femto hotspots in macro cells.


Figure 3-2. Benefits of X2 broker: scalability and stability

3.3 X2 Service Broker Operation

In order for X2 service broker to work, HeMS allocates HeNB IDs to HeNBs as seen in Figure 3-3. An HeNB ID is 28 bits long, and consists of i) an eNB ID (20 bits long), identical for all HeNBs (up to 256) that belong to the same virtual eNB, and ii) a cell ID (8 bits long), unique for all the HeNBs (up to 256). This HeNB ID plaNning scheme lets a macro eNB recognize a virtual eNB as just another macro eNB, and all the HeNBs belonging to it as its cells.


Figure 3-3. SMEC X2 service broker: HeNB ID planning

Detailed call flow for X2 broker operation is as follows:

❶ HeNB1 initiates TNL address discovery procedure towards an MeNB: HeNB1 detects a new cell (cell A of macro eNB) and decides to setup X2 towards Macro eNB (MeNB). It initiates an TNL address discovery procedure by sending eNB Configuration Transfer message indicating its own HeNB ID (HeNB1, 28 bits long) and MeNB ID (20 bits long) as neighbor information to virtual eNB through S1 interface.

The virtual eNB does not have any information on the MeNB’s X2 IP address, and it must forward the message to MME to find the X2 IP address of MeNB. Before forwarding the message, virtual eNB (X2 broker) replaces the 28-bit HeNB ID with its own ID (virtual eNB, 20 bit long) in the message and forwards it to MME. MME knows the MeNB and so sends an MME Configuration Transfer message to it (note that virtual eNB does not disclose 28-bit-long HeNB ID to MME and MeNB).


Figure 3-4. SMEC X2 service broker: HeNB1 initiates TNL address discovery procedure towards an MeNB

MeNB returns its X2 IP address, and MME sends it to virtual eNB (now, virtual eNB obtains MeNB’s X2 IP address). Virtual eNB replaces the MeNB’s X2 IP address in SeNB Information with its own IP address, and sends MME Configuration Transfer message to HeNB1. Then, this leads HeNB1 to recognize the virtual eNB IP address as MeNB’s X2 IP address.

❷ X2 setup between HeNB1 and MeNB: HeNB1 starts X2 setup towards MeNB, indicating its HeNB ID (virtual eNB (20b) + cell 1 (8b)) and MeNB as neighbor information. Since HeNB1 knows virtual eNB’s IP address as MeNB’s X2 IP address, this message is actually forwarded to virtual eNB. Virtual eNB starts another X2 setup procedure to continue the setup of X2-connectivity towards MeNB, indicating its own eNB ID (virtual eNB) and cell information (cell 1) and MeNB ID as neighbor information. When MeNB and virtual eNB responds, a single X2 connection is set up between HeNB1 and virtual eNB, and also between virtual eNB and MeNB.
This process lets MeNB add the cell information of HeNB1 (virtual eNB/cell1) to its X2 neighbor list and also lets HeNB1 add the cell information of MeNB (MeNB/cell A) to its X2 neighbor list.

Figure 3-5. SMEC X2 service broker: X2 setup between HeNB1 and MeNB

❸ Subsequent X2 connection setups: As X2 connection between virtual eNB and MeNB has already been setup, any further X2-address request from other HeNBs for X2-connectivity towards MeNB will be responded by the virtual eNB without forwarding the request via the MME towards the MeNB. Virtual eNB sends its own IP address in response to other HeNB’s X2-address request to the MeNB.

For any further X2 setup request to the MeNB, virtual eNB, through the already-established X2 connection, sends an X2 message (eNB Configuration Update) containing HeNB2 cell information to inform MeNB of the updated cell information.

Virtual eNB sends X2 Setup Response to HeNB2 if the X2 Configuration Update between the virtual eNB and MeNB is performed successfully.

Figure 3-6. SMEC X2 service broker: Subsequent X2 connection setups

Once the above process is completed, an X2 connection is set up between each HeNB and virtual eNB (HeNB-GW), and also between virtual eNB and macro eNB. Logically, existing macro eNB recognizes HeNB-GW as a new macro eNB, and all HeNBs belonging to it as cells in the macro eNB as shown in Figure 3-7.


Figure 3-7. SMEC X2 service broker: Logical configuration

This means, the legacy LTE network (eNB and EPC) will see even a large-scale deployment of HeNBs as a small-scale deployment of additional macro eNBs. This completely eliminates any chance of uncertainty, complexity, or risk factors that would otherwise be caused by a large-scale deployment of HeNB in the legacy LTE network. For example, because the 28-bit HeNB IDs are not exposed to MME or eNB, there is no potential issue in interworking between HeNBs and MME/eNBs, which makes the network architecture even more stable and reliable.

As the X2 service broker feature by SMEC is implemented using S1 interface (eNB n MME) and X2 interface (eNB n eNB) defined in Rel. 8, no change or modification is needed in the EPC core or eNBs already deployed in the legacy LTE network. This makes the feature readily applicable to any LTE commercial network where Rel. 8 or higher is implemented (i.e., in any LTE network).


4. Benefits of SMEC HeNB-GW

SMEC’s HeNB-GW helps to keep the impact of introducing LTE femtocell – even when massively deployed – in the legacy LTE network low, as low as that of small scale addition of macro eNB. This ensures the stability of the LTE core network remains unaffected and the additional investment costs resulting from such deployment are kept to a minimum.

  • SMEC’s HeNB-GW delivers both SeGW feature and aggregation feature (for control plane, S1-MME and user plane, S1-U) at a single point, proactively preventing overload at existing MME and S-GW, and also easing potential uncertainty in the legacy LTE network to be caused by tens of thousands of newly deployed femtocells. Also, it helps to bring down the costs for additional installation of MME resulting from the large scale deployment of femtocells (e.g. purchasing additional equipment and license).
  • SMEC’s HeNB-GW supports S1 and X2 handover between macro eNB and femtocell, which ensures uninterrupted, reliable call quality, even during switches between the two cells – all just through 4G network (i.e. just through VoLTE) without 2G or 3G.
  • SMEC’s HeNB-GW offers X2 service broker feature that provides X2 handover between macro eNB and HeNB without having to modify X2 interface used between the macro eNBs.
  1. Traditional HeNB-GW can only support S1 handover, and thus heavy overloads are inevitably passed on to MME during handover. SMEC HeNB GW, however, supports X2 handover where no MME involvement during handover process is needed, drastically reducing overload at MME.
  2. It significantly reduces the number of X2 interfaces needed through aggregation of X2 interfaces between macro eNB and femtocells, thereby decreasing network complexity to be caused by X2 interface used in small cell environment.
  3. The X2 service broker feature by SMEC, implementable through S1 and X2 interfaces defined in 3GPP Rel. 8., is readily deployable in any LTE system regardless of its release version. Without additional installation of X2 GW nodes defined in R-12 or upgrade of R-12 X2 GW feature license of MME, eNB and HeNB, or of LTE network, X2 handover between macro eNB and HeNB can be readily supported.



3GPP 3rd Generation Partnership Project
eNB Evolved Node B
EPC Evolved Packet Core
GTP GPRS Tunneling Protocol
GW Gateway
HeMS HeNB Management System
HeNB Home eNodeB (Femtocell)
HeNB-GW Home eNodeB Gateway (Femto Gateway)
ID Identifier
IMS IP Multimedia Subsystem
ISP Internet Service Provider
LTE Long Term Evolution
MeNB Macro eNB
MME Mobility Management Entity
PGW Packet Data Network Gateway
QoE Quality of Experience
RAN Radio Access Network
SCTP Stream Control Transmission Protocol
SeGW Security Gateway
SeNB Source eNB
SOHO Small Office Home Office
S-GW Serving Gateway
TNL Transport Network Layer
UE User Equipment
VoLTE Voice over LTE
X2 AP X2 Application Protocol
X2 GW X2 Gateway


SMEC’s LTE Femto Gateway with X2 Broker – Facilitating instant mass deployment of LTE femtocells in existing LTE infrastructure
March 14, 2016 | By Y.C. Lee (www.esmec.com), Dr. Harrison Jangwoo Son (tech@netmanias.com)


Source: http://www.netmanias.com/en/?m=view&id=reports&no=8493


How to manage the LTE revolution in Asia-Pacific with next generation backhaul

2 Oct

LTE growth is being driven by consumer demand for data, the absence of fixed line infrastructure in many parts of emerging APAC (EMAP), and the need to provide the network capacity to enable next-generation mobile and services.

Operators are desperately looking to efficiently scale network capacity; wireless technology holds the key to delivering the performance and profits operators require as the mobile landscape changes dramatically.

Full Article

Consumers all over the world want the fastest network, with the highest quality of experience. This is no more evident than in the Asia-Pacific (APAC) region where LTE is now out of the experimental stage and being deployed widely across most of the developed markets in the region. According to a new report by Allied Market Research, APAC is forecast to surpass other geographical markets by 2020 with approximately 40 per cent of the global LTE market. Analysys Mason is also forecasting that APAC and Latin America will account for the majority of the networks that are planned for launch by 2018. A recent report from the Global Mobile Suppliers Association confirms the demand for LTE networks, estimating an approximate 200 million LTE subscribers globally with the APAC region boasting 77.8 million, a 38.8 per cent share of the overall subscribers.

LTE growth drivers

LTE growth is being driven by consumer demand for data, the absence of fixed line infrastructure in many parts of emerging APAC (EMAP), and the need to provide the network capacity to enable next-generation mobile and services. Rapid economic development, which has increased the region’s prosperity, has also been a factor in making mobile services more affordable and helped seed the LTE ambitions of operators.

In addition, access to high-speed LTE is facilitating a wide variety of socio-economic benefits across APAC, encouraging governments to incentivise operators to deploy next generation networks. LTE is helping people lead more productive lives and, for example, enabling businesses to become more efficient in delivering goods and services. The onset of widespread broadband connectivity across the region is sustaining this economic development with improved networks in some of the countries in the EMAP region, empowering education, increasing trade and driving innovation.

LTE diversity

Growth in LTE, and the subsequent rise in mobile data traffic, is leading to an increase in infrastructure investment. Operators have the challenge of efficiently scaling infrastructure which delivers the capacity to satisfy consumer appetite for mobile connectivity and support the array of new services being deployed across the region.

This challenge is evident in the diversity of development across APAC’s mobile market which has led to a multitude of LTE network adoption scenarios. The variety is evident in the 47 countries and 3.7 billon people in the region which contain many intricacies and complexities due to economic, political and geographic factors. South Asia, for instance, reflects a diverse mix of mobile and internet diffusion patterns. Malaysia and Singapore have a mature network infrastructure and mobile penetration exceeding 100 per cent, whilst countries like the Philippines and Indonesia are still considered to have a developing infrastructure.

It is expected that these EMAP regions will be able to take most advantage of the demand for LTE networks rather than the developed APAC (DVAP) regions that have more mature offerings. However, managing and constructing an LTE network has many factors to consider, not least the technical requirements needed for mobile backhaul. As always, the cost of backhaul is a paramount consideration in running and launching new networks.

The Philippines is a good example of the complexities of managing LTE networks. A recent report by OpenSignal Inc. has concluded that the Philippines have the slowest LTE connection among the 16 countries surveyed, with 5.3 Mbps (megabits per second). Operators in EMAP regions have increasing pressure to provide the capacity needed to handle the huge data demands from smartphones, tablets and new technologies such as M2M. Operators are in danger of failing to provide consumers and business with the fast, high quality network that is demanded of LTE.

Wireless innovation

Operators have known for some time that they need to drive innovation in their business processes and run networks at a much lower cost per bit to achieve success. However, the extensive capital expenditures (CAPEX) and operating expenditure (OPEX) challenge in setting up new infrastructure is seeing operators struggle to make a successful business case. For example, putting vast amounts of fibre networks into the ground can encounter huge costs and lengthy time to the market as well as a geopolitical minefield of regulation which can reduce an operator’s return on investment (ROI). Even worse, fibre can suffer from poor reliability and high maintenance costs due to either deliberate or accidental damage. Increasingly, operators are turning to a new wave of efficient, flexible and high capacity wireless technologies, including point-to-multipoint (PMP) microwave.

Traditionally, the low ARPU in the APAC countries puts even more emphasis on operators to make efficiencies. This means that, for the overall business case to work, every bit of data must be delivered at the lowest possible cost, and it’s this imperative that makes operators turn to innovative solutions like PMP microwave. Because the hub radio itself, as well as the backhaul spectrum, are shared across a number of LTE sites in the sector; both the hub equipment and spectrum cost are amortised across this number of links. Analyst consultancy Senza Fili recently found this allows PMP microwave to deliver savings of up to 50 per cent over other forms of backhaul, while delivering the same carrier-grade service essential for LTE.

PMP microwave uses area-licensed spectrum to create a sector of backhaul coverage from a single hub site and ensures the guaranteed quality of service LTE demands. Multiple cell sites can be backhauled within this sector, and bandwidth is dynamically shared across all links. Due to this real-time allocation of spectrum, PMP microwave enables the ‘troughs’ of one cell site’s traffic demands to be filled by the ‘peaks’ of another. This aggregation reduces the total bandwidth required for a sector and has been proven to improve spectral efficiency by at least 40 per cent when compared to traditional point-to-point (PTP) technology. By efficiently managing the backhaul spectrum required for LTE, operators can run networks at a much lower cost and achieve a higher ROI – crucial at a time where revenues are under threat. Importantly, PMP microwave (which operates above 6GHz) has the capacity to handle the most demanding LTE networks and is already proven in LTE backhaul deployments in other regions of the world.

Enterprise access

Whilst the enormous promise of LTE is clearly evident, operators still need to look at new and innovative ways to unlock the true potential of their backhaul infrastructure and increase ROI. Many operators see the deployment of multiple virtual networks over a common physical network as the answer. Some operators currently choose to build completely new LTE or enterprise access networks to sit alongside legacy infrastructure, however this can create inefficiencies across the different generations of technologies.

The latest backhaul technology now allows for new profitable business models to be created. By creating a converged backhaul network, LTE backhaul can be accommodated whilst also using the virtual networking capability to monetise spare capacity by deploying additional services to businesses. A converged PMP microwave backhaul network, for instance, enables operators to introduce fixed enterprise access services on the same LTE network – serving business with next generation connectivity.

This efficient use of backhaul and spectrum enables operators to invest in fast mobile speeds and carrier grade services, whilst allowing for competitive pricing and increasing profitability. This increase in ROI is particularly beneficial at a time where the fragmentation of spectrum is a particular issue for APAC.


The long term growth prospects for mobile broadband in APAC are enormous as operators are finding consumers and businesses hungry for transformational mobile and internet services. With operators desperately looking to efficiently scale network capacity, wireless technology holds the key to delivering the performance and profits operators require as the mobile landscape changes dramatically.

New business models and innovative wireless backhaul will not only protect investments in LTE but pave the way for new services and revenue opportunities – helping operators reduce churn in what is becoming an increasingly competitive market.

It is an exciting opportunity for operators in APAC to upgrade their technology for LTE and bring new innovative services to the market. With cost savings obtained by increased efficiency and utilisation of resources, quality of service or features need not be sacrificed with wireless technologies. As customer preferences change and mature in the APAC region, there is huge potential in the market to deploy efficient and flexible wireless technologies to build fast successful networks.

Source: http://www.connect-world.com/index.php/magazines/asia-pacific/item/24555-how-to-manage-the-lte-revolution-in-asia-pacific-with-next-generation-backhaul

LTE Asia: transition from technology to value… or die

27 Sep


I am just back from LTE Asia in Singapore, where I chaired the track on Network Optimization. The show was well attended with over 900 people by Informa’s estimate.

Once again, I am a bit surprised and disappointed by the gap between operators and vendors’ discourse.

By and large, operators who came (SK, KDDI, KT, Chungwha, HKCSL, Telkomsel, Indosat to name but a few) had excellent presentations on their past successes and current challenges, highlighting the need for new revenue models, a new content (particularly video) value chain and better customer engagement.

Vendors of all stripes seem to consistently miss the message and try to push technology when their customer need value. I appreciate that the transition is difficult and as I was reflecting with a vendor’s executive at the show, selling technology feels somewhat safer and easier than value.
But, as many operators are finding out in their home turf, their consumers do not care much about technology any more. It is about brand, service, image and value that OTT service providers are winning consumers mind share. Here lies the risk and opportunity. Operators need help to evolve and re invent the mobile value chain.

The value proposition of vendors must evolve towards solutions such as intelligent roaming, 2-way business models with content providers, service type prioritization (messaging, social, video, entertainment, sports…), bundling and charging…

At the heart of this necessary revolution is something that makes many uneasy. DPI and traffic classification, relying on ports and protocols is the basis of today’s traffic management and is becoming rapidly obsolete. A new generation of traffic management engines is needed. The ability to recognize content and service types at a granular level is key. How can the mobile industry can evolve in the OTT world if operators are not able to recognize a content that is user-generated vs. Hollywood? How can operators monetize video if they cannot detect, recognize, prioritize, assure advertising content?

Operators have some key assets, though. Last mile delivery, accurate customer demographics, billing relationship and location must be leveraged. YouTube knows whether you are on iPad or laptop but not necessarily whether your cellular interface is 3G, HSPA, LTE… they certainly can’t see whether a user’s poor connection is the result of network congestion, spectrum interference, distance from the cell tower or throttling because the user exceeds its data allowance… There is value there, if operators are ready to transform themselves and their organization to harvest and sell value, not access…

Opportunities are many. Vendors who continue to sell SIP, IMS, VoLTE, Diameter and their next generation hip equivalent LTE Adavanced, 5G, cloud, NFV… will miss the point. None of these are of interest for the consumer. Even if the operator insist on buying or talking about technology, services and value will be key to success… unless you are planning to be an M2M operator, but that is a story for another time.

%d bloggers like this: