2.5.1 With or Without an Anchor in the Licensed Band
As described in Section III.A, there are two deployment scenarios with respect to whether the licensed and unlicensed spectrum are aggregated or not. The choice between the two deployment options is dictated by the operator’s existing assets and deployment plans.
Aggregating licensed and unlicensed spectrum for operator-controlled access to un- licensed spectrum that is well integrated to the LTE core network [3] can offer signifi- cant advantages. First, aggregating licensed and unlicensed bands can enable operators to leverage the existing LTE hardware in both the radio and core networks, thus, data offloading can be achieved in a seamless fashion. Moreover, to manage the different component carriers, the LTE eNB operating in the licensed spectrum can carry the con- trol signaling which is granted the highest priority among nine Quality-of-Service (QoS) class identifiers defined by LTE. The signaling and control information is crucial not only to ensure the resource allocation is managed properly but also to maintain the ro- bustness of the links. In some cases where there is a lot of interference and all the nodes are competing for resources, it is crucial to allocate the resources with an order. In addi- tion, LTE macro cell can provide ubiquitous coverage for UEs. The above features make LTE able to facilitate opportunistic unlicensed access.
However, technologies requiring the operator to have an anchor in licensed spectrum may in practice limit the potential uses. A MulteFire network which is operating entirely in unlicensed spectrum seems to be promising for small businesses, enterprises, venue owners, and cable operators who lack licensed spectrum.
2.5.2 Radio Link-Level or TCP-level LTE-WLAN Integration
In radio link-level integration, LTE and Wi-Fi networks are closely coupled to potentially provide the highest performance, but with high implementation complexity. Compared to the TCP-level integration solution, where the transport layer has to infer the conges-
tion in the links using round trip delays and TCP acknowledgement loss, LWA and LWIP can improve system performance by managing radio resources in real time according to the radio frequency (RF) and load conditions of both LTE and Wi-Fi.
MPTCP proxy performs aggregation on top of the legacy networks, thus requiring no change in the legacy networks. A drawback to the radio link-level integration solution is investment costs for replacing less capable eNBs and Wi-Fi APs. On the contrary, LWPA based on MPTCP requires only modifications to operate software in the client devices and servers, and hence is easily implementable. It is also compatible with any legacy Wi-Fi APs. Since LTE and Wi-Fi have their own networks, the MPTCP proxy should identify ways to perform flow control on traffic forwarded to the respective network.
2.5.3 Comparison of Standardization Statuses
3GPP is making efforts to standardize LTE-LAA, LWA and LWIP in order to increase LTE rate through leveraging unlicensed Wi-Fi bands. In particular, LTE-LAA was ap- proved as a Work Item (WI) for Release 13 in June 2015. 3GPP specified LTE-LAA for DL operation in Release 13 and is currently working on specifying LTE-LAA for UL operation in Release 14. Dual connectivity supporting spectrum aggregation be- tween macro and small cells is another important LTE-LAA feature expected in Release 14 and beyond. LTE-LAA is standardized as a single global solution to be adopted by all regions with or without LBT requirements. The standardization process for LWA, including protocol architecture, solutions for aggregating data at the PDCP layer, sig- naling and interfaces between eNB and Wi-Fi AP, etc., was completed in March 2016. 3GPP also works on the enhanced LWA (eLWA) in Release 14. The eLWA is built on the Release 13 LWA framework without any change on the LWA architecture. Main topics include UL support, enhanced mobility, support for 60 GHz, and optimizations for high data rates 802.11 technologies, i.e., 802.11ax, 802.11ad and 802.11ay. Standardization for LWIP was formally completed in March 2016. The mobile operators can implement DL and UL functionality right away with LWIP.
The LTE-U forum has released LTE-U specifications including duty-cycle fair access solution since 2015. The LTE-U is a non-standard technology that employs a proprietary coexistence algorithm. The MulteFire Alliance released the MulteFire specification in 2016, which is built on elements of 3GPP Release 13 LTE-LAA for the DL and Release 14 enhanced LTE-LAA for the UL. Although standardization for MPTCP proxy began in
2009 when Internet Engineering Task Fore (IETF) MPTCP working group was formed, detailed architecture and deployment scenarios of MPTCP proxy-based aggregation, i.e., LWPA, have not been specified yet.
2.5.4 Further Research Directions
To meet the LTE and Wi-Fi inter-working challenges proposed in section 1.2.3, the com- munity has proposed several coexistence mechanisms for both markets with and without LBT requirement. However, as will be discussed in the following sections, such a kind of coexistence is not going smoothly thus far. I summarize the key challenges related to the LTE and Wi-Fi cooperation as follows:
a) Disputes over the effectiveness on current coexistence mechanisms are still the
hot topic of the community.As will be discussed in section 2.3.4, both duty-cycling
and LTE-LAA with LBT are designed for specific markets. What’s more, as stated in section 2.3.1 and 2.3.4, both mechanisms have their own weaknesses, and dispute remains over whether these mechanisms are valid in some specific scenarios. An agreement among the community is needed on one or more acceptable coexistence mechanisms.
b) The lack of documented agreement on a definition of fairness is a big problem. As stated in section 1.2.3, there exist different kinds of definition of fairness. The situation that Wi-Fi stakeholders tend to accept that fairness criteria means LTE- LAA should not impact Wi-Fi more enormous than another Wi-Fi network. Some 3GPP members believe that fair access means that LTE-LAA BS and IEEE 802.11 clients should have half of the bandwidth respectively. An agreement is required on the definition of fairness or a mechanism that achieves fairness.
c) It is still too early to determine whether LTE-LAA is successful. LTE-LAA is just one of a number of spectrum-sharing methods being used now with others in develop- ment or in test trials. The disagreement among different members in the community shows that there is no unified test platform. Furthermore, more researches concerning coexistence optimization are required. For example, new objective functions for op- timizations problem formulations to guarantee the fair coexistence of LTE-LAA and Wi-Fi are needed. The attributes of the CR to optimize LTE-LAA in 5 GHz are also required. What’s more, more complex tests on fairness, especially those based on a
range of realistic usage scenarios are urgently needed. That means, before drawing any conclusions, the community should first complete simulations representing more realistic usage scenarios.
d) It lies in the features of specific scenarios that decide the coexistence is necessary
or unnecessary. That is to say, if it is not worthwhile for operators to deploy LTE-
LAA from the perspective of various performance metrics, the coexistence is not necessary accordingly. From the view of market and technology, both LTE-LAA and Wi-Fi have their own benefits and cannot be replaced by each other. In this case, the choice of Wi-Fi or LTE-LAA is also related to the experience of operators and even financial and other factors.