CUPS – Set to be a key core network feature for many operators. 3GPP Release 14.

Control and User Plane Separation of EPC nodes (CUPS). 

CUPS being the 3GPP Relase 14 standards, but most fundamental aspect in evolution of Next Generation Networks Architecture, As we since last 3-4 years have been in continuous support of complete separation of control and user plan for bringing capability in network for required flexibility in terms of service creation and required scalability.

Though all IP paradigm shif since the inception of 4G has taken the first step towards it but recent advancement on virtualisation like of cloud and NFV has taken the troll ahead. 

in terms of 3GPP word..

CUPS stands for Control and User Plane Separation of EPC nodes and provides the architecture enhancements for the separation of functionality in the Evolved Packet Core’s SGW, PGW and TDF. This enables flexible network deployment and operation, by distributed or centralized deployment and the independent scaling between control plane and user plane functions - while not affecting the functionality of the existing nodes subject to this split.

CUPS allows for:

• Reducing Latency on application service, e.g. by selecting User plane nodes which are closer to the RAN or more appropriate for the intended UE usage type without increasing the number of control plane nodes.
• Supporting Increase of Data Traffic, by enabling to add user plane nodes without changing the number of SGW-C, PGW-C and TDF-C in the network.
• Locating and Scaling the CP and UP resources of the EPC nodes independently.
• Independent evolution of the CP and UP functions.
• Enabling Software Defined Networking to deliver user plane data more efficiently.

Architecture principles (Working Group SA2)

CUPS introduces 3 new interfaces, Sxa, Sxb and Sxc between the CP and UP functions of the SGW, PGW and TDF respectively. 

 

The following high-level principles were adopted:

• The CP function terminates the Control Plane protocols: GTP-C, Diameter (Gx, Gy, Gz).
• A CP function can interface multiple UP functions, and a UP function can be shared by multiple CP functions.
• An UE is served by a single SGW-CP but multiple SGW-UPs can be selected for different PDN connections. A user plane data packet may traverse multiple UP functions.
• The CP function controls the processing of the packets in the UP function by provisioning a set of rules in Sx sessions, i.e. Packet Detection Rules for packets inspection, Forwarding Action Rules for packets handling (e.g. forward, duplicate, buffer, drop), Qos Enforcement Rules to enforce QoS policing on the packets, Usage Reporting Rules for measuring the traffic usage.
• All the 3GPP features impacting the UP function (PCC, Charging, Lawful Interception, etc) are supported, while the UP function is designed as much as possible 3GPP agnostic. For example, the UPF is not aware of bearer concept.
• Charging and Usage Monitoring are supported by instructing the UP function to measure and report traffic usage, using Usage Reporting Rule(s). No impact is expected to OFCS, OCS and the PCRF.
• The CP or UP function is responsible for GTP-u F-TEID allocation.
• A legacy SGW, PGW and TDF can be replaced by a split node without effecting connected legacy nodes.

Protocol Design (Working Group CT4)

CT4 assessed candidate protocols such as OpenFlow, FoRCES, Diameter, IETF DMM FPC and 3GPP native protocol. The criteria identified for the selection process included;

• ease of implementation on simple forwarding devices,
• no transport Head-Of-Line blocking,
• low latency,
• capabilities to support all the existing 3GPP features
• ease to extend and maintain the protocols to support 3GPP features,
• backward compatibility across releases

Based on these criteria, CT4 decided to define a 3GPP native protocol with TLV encoded messages over UDP/IP, called Packet Forwarding Control Plane (PFCP) protocol, for the Sxa, Sxb and Sxc interfaces.

 

Details on the protocol on Sx

PFCP has the following main properties:

• One Sx Association shall be setup between a CP function and a UP function before being able to establish Sx sessions on the UP function. The Sx association may be established by the CP function (mandatory support) or by the UP function (optional support).
• An Sx session is established in the UP function to provision rules instructing the UP function how to process a certain traffic. An Sx Session may correspond to an individual PDN connection, TDF session or this can be a standalone session not tied to any PDN connection/TDF session, e.g. for forwarding DHCP/RADIUS/DIAMETER signalling between the PGW-C and PDN (SGi).
• Sx Node related procedures:
  - Sx Association Setup / Update / Release procedures;
  - Heartbeat procedure to check that a PFCP peer is alive;
  - Load Control and Overload Control procedures to balance the load across UP functions and reduce signalling towards UP function in overload;
  - Sx PFD Management procedure to provision PFDs (Packet Flow Descriptions) for one or more Application Identifiers in the UP function (SDCI).
• Sx Session related procedures:
  - Sx Session Establishment / Modification / Deletion procedures;
  - Sx Session Report procedure to report traffic usage or specific events (e.g. arrival of a DL data packet, start of an application).
• Data Forwarding between the CP and UP functions is supported by GTP-U encapsulation, e.g. for forwarding RS/RA/DHCP signalling between UE and PGW-C, or forwarding user plane data to the SGW-C when buffering of DL packets is done in the CP function. 
• PFCP supports reliable delivery of messages.
• New DNS procedures are defined for UP function selection. The CP function selects a UP function based on DNS or local configuration, the capabilities of the UP function and the overload control information provided by the UP function.

Restoration procedures

CT4 has analysed the possible error scenarios with split EPC nodes and defined corresponding restoration solutions in TS 23.007.

Further Reading:

Stage 2 – Functional Architecture and Procedures

• TS 23.214 Architecture enhancements for control and user plane separation of EPC nodes.

Stage 3 – Protocols

• TS 29.244 Interface between the Control Plane and the User Plane of EPC Nodes.
• TS 23.007 Restoration procedures
• TS 29.303 DNS procedures for UP function selection  

Antenna theory, basics in modern technological perspective

Antenna basics from-r&s from Saurabh Verma

How to make 5G an energy efficient system? by Marcin Dryjanski, Grandmetric.

One of the key requirements for 5G is the reduction of network energy consumption. Example techniques to assure high energy efficiency in mobile networks include: the use of lean carrier, dynamic beamforming, efficient energy saving schemes for Small Cells, virtualization of the network and utilization of dense deployments.

Energy efficiency for 5G networks

As far as I know, there are no strict values defined in any recommendations yet regarding the specific target energy level in bits/Hz/J. However, the general target is that the energy consumption for 5G should not be larger than that of today’s networks (IMT [1]), or larger than half of what the current networks consume (NGMN [2]), while still supporting 1000x capacity increase. This means the energy efficiency of 5G systems should increase by a factor of 1000x (IMT) or 2000x (NGMN). 3GPP also wants to evaluate and compare different solutions first [3].

http://www.xgnlab.com

Energy efficiency in the network can be achieved by techniques that span from PHY layer to up to networking and architecture levels. Below a subset of those is presented.

• Lean carrier design – a radio frame without (or with limited) “always-on” signals like Cell-specific Reference Signals. This approach aims at minimization of the transmission that is not directly related to data transaction. On one hand it provides more time without any transmission (at the network side) and on the other hand the UE can also be longer in a sleep mode.

• Dynamic beamforming – utilization of advanced beamforming and MIMO schemes enable focusing most of the transmit energy towards a specific receiver (localized transmission), which in turn decreases the energy being dissipated over an area where there are no users, thus saving this energy.

• Dynamic energy saving schemes for small cells – a concept where the network nodes transmit only when and where needed. This enables utilization of the cell being OFF in the times when there are no users to serve (according to real life measurements most cells are “empty” for a significant part of the day). An example could be an energy saving scheme with two level sleeping: deep sleep, where no synchronization is possible; light sleep, in which discovery of the cell is possible and access is possible (cell transmits only necessary signals); and in normal operation mode – enable data traffic transactions. Additionally, in this CP/UP split may be utilized, where macro sites provide control plane connectivity and the small cells are used as pure UP nodes (i.e. capacity/data boosters that can be switched on only upon need).

• Densification of the network with very small cells – bringing more and more nodes to the network and reducing the ISD and distance to a user (i.e. deploying them below rooftop) reduces the pathloss and thus the necessary power to reach a user (and this in turn decreases the energy utilization at the UE side and required energy at the individual network node).

• Network virtualization and advanced network sharing – utilization of single infrastructure by multiple operators, especially for dense networks, described just above, with multitude of small cells, divides the energy consumption for each participating operator. Additionally, the network slicing concept provides the possibility to utilize only the necessary functions that are tailored and optimized for a specific application. These functions can be flexibly assigned to the processing nodes, thus can be deployed in the most efficient “places” with respect to e.g., energy efficiency. Additionally, the hardware can be utilized more efficiently benefiting from pooling gains.

For more insight go to source HERE.

Blog Author:

MARCIN DRYJANSKI

Marcin Dryjanski received his M.Sc. degree in telecommunications from the Poznan University of Technology in Poland in June 2008. During the past 8 years, Marcin has served as R&D Engineer, Lead Researcher, R&D Consultant, Technical Trainer and Technical Leader. He has been providing expert level courses in the area of LTE/LTE-Advanced for leading mobile operators and vendors. Marcin provides consulting services to business projects in the area of 5G related topics. In addition to that, Marcin was a workpackage leader in EU-funded research projects aiming at radio interface design for 5G including FP-7 5GNOW and FP-7 SOLDER. He co-authored a number of research papers targeting 5G radio interface design. To contact Marcin please write to: marcin.dryjanski@grandmetric.com

NGMN : 5G E2E Architecture framework; more on 'large scale convergence' at data plane.

The 5G system will also support flexible RAN structures including implementations based on Cloud principles and the placement of context awareness at the RAN edges (i.e. mobile edge computing). Both centralized and distributed implementation of RAN functions should be enabled to facilitate the realization of various RAN implementations. In addition, support for various coverage layers and cell sizes spanning extreme long-distance covering macro cells to small cell radio access deployments is required.

An open Perspective for LARGE SCALE CONVERGENCE, specially in case of data plane.

RAN Decomposition, Functional decomposition of the radio network is required to meet the diverse information transport demands (high performance to low performance) and align them with the demands of next-generation service categories of eMBB, mIoT, and URLLC. To accommodate these, a decomposition of the radio network protocol layer functions, across layer-1, layer-2, and layer-3 is required, in terms of the degree of centralisation or distribution.

This decomposition consists of placing more functions of the upper layers of the radio network protocol stack in distributed entities for high performance transport demands (e.g. high bandwidth, high-capacity, low-latency, low jitter etc.,) relative to a centralized entity. Scheduling optimisation at a centralized entity, for high performance transport across multiple distributed entities (e.g. base stations, remote radio heads etc.) for fast coordination is critical requirement.

For relatively low performance transport, more of the upper layer of the radio network protocol stack is placed at a centralized entity to optimize the cost/performance trade-off, associated with the distributed entities. This choice of functional split will determine the x-haul capacity requirement and associated latency specifications and performance. This will impact the network architecture as it could determine the placement of nodes and distance between them or, in the case of a higher layer split, will be tolerant of a large latency from a RAN perspective which may be excessive when low-latency services are considered, therefore bounds must be applied within the network architecture to enable a service provider to support low latency services.

A distributed RAN (D-RAN) with several functional splits will be supported by 5G. Figure 1 illustrates the configuration with co-located centralised unit (CU) and distributed (DU). All radio protocol layers are terminated within the cell site.

The connection from the cell site towards the core network is traditional mobile backhaul which will be scaled and optimised to support 5G data rates and performance targets such as low-latency, low PELR, low and very deterministic PDV etc. The D-RAN configuration does not constrain the ability of the local CU to support remote DU; in fact the cell site could become a CU for other cells sub-tended as illustrated in Figure 2.

A 5G C-RAN can be implemented with a higher layer split with the protocols stack with PDCP being located in the CU while the remainder of the stack is in the DU, as shown in Figure 3. This is one example; other splits will result in a different distribution of functionality between CU and DU.

This configuration has similar x-haul capacity requirements when compared with traditional backhaul, the latency and performance requirements of the RAN are not stringent and therefore consideration must be given to engineer the x-haul link in accordance with service-based latency and performance targets.

170511 ngmn e2_e_archframework_v0.6.5 from Saurabh Verma

WiFi & LTE Convergence at RAN level using 3GPP LIMONET standards.

This is in coninuation, rather to say,  to supplement  my earlier blog on WiFi offload or rightly saying WiFi & LTE convergence. putting down some excerpt from that     ......."http://blog.fundarc.co.uk/2016/09/wifi-is-much-required-to-offload-many.html"

What is of interest....

Wi-Fi is not there for hand off or to say in popular term offload. It could be leveraged as associated network. An associated data path to LTE network by converging at RAN or IP-RAN through a Local PGW and keeping the common EPC core. This convergence is possible if S1AP will be terminated at this Local PGW and MME is kept transparent to converged Local network. EPC/PGW while creating the bearer, based on IPCAN session (contains the policy), will provide the information, based on policy, which will help Local PGW to distinguish and forward the data flow.

Wi-Fi and cellular could be converged at radio access level and Wi-Fi can be used an associated data channel for the mobile access. This feasibility is well accepted now and companies like Qualcomm have more obsessed approches. Qualcomm is pursuing the aggregation at radio link level with a point of convergence at PDCP layer (https://www.qualcomm.com/invention/research/projects/lte-advanced/lte-wi-fi-interworking).

On the convergence of two at RAN, the small cell forum is also putting its attention through research and industry surveys. Its recent whitepaper, feb 2016, titled "Industry perspectives, trusted WLAN architectures and deployment considerations for integrated Small-Cell Wi-Fi (ISW) networks" SCF states at section 2.0 Integrated small cell Wi-Fi (ISW) networks..... "other interesting alternatives are possible, namely integration in the SC-APs (i.e. RAN-based integration) and/or in SC-gateways (i.e. GW-based integration). Here, the Integration function resides at the edge, possibly in an integrated ISW-AP. RAN-based integration of licensed/unlicensed access is now being addressed by 3GPP release 13, including approaches for RAN based integration of Wi-Fi and LTE.

Finally, architectures that integrate Wi-Fi and SCs at the gateway level are possible. For example, the SC-GW (i.e. H(e)NB-GW) as well as Wi-Fi GW (i.e. ePDG and/or TWAG/TWAP) may be realized together, along with associated integration functions. At the time of writing, these architectures are still in consideration and development."

So that's the future.....

This could be done at network level, where you need not to go all the way to PGW at EPC but having a local PGW at RAN/IP-RAN network. The advantage it will bring is the common EPC for both cellular and Wi-Fi networks i.e. no need for the transfer of PDN connectivity context like an inter-RAT scenarios. Only thing is that the solution is not for any kind of Wi-Fi device but it is for cellular device equipped with Wi-Fi as we need cellular for all the service control level functionality.

Service delivery is well controlled through the 3GPP PCC architecture that define PCRF, PCEF (PGW) and Application server (AS) level interaction and coordination also referred to as IPCAN session. In EPC we need bearer to carry the traffic for specific service. This bearer traffic is delivered to RAN to reach to end devices over air interfaces or radio channels. EPC provide all the necessary information as QoS parameter for specific bearer to RAN for required radio channel capacity. The interface between EPC and RAN is S1 (S1-AP for control path and S1-U for data path).

Instead of having direct interface between eNB and EPC if we keep a Local PGW node (L-PGW) to interface with EPC and provide convergence of Wi-Fi network at this L-PGW . A replica of this PCC architecture can be implemented at this L-PGW level which will decide to deliver the service at Wi-Fi of Cellular radio. This will definitely need the modification on RAN and EPC interfaces like s1AP and NAS etc. as L-PGW has to coordinate with MME.

There is nothing to scratch to get go, the provisions of such nature framework and already there with 3GPP, like for LIMONET (http://www.3gpp.org/DynaReport/WiVsSpec--500028.htm). We can leverage this for convergence of the two networks at local network level with a common core. The Standards for LIPA specific work at 3GPP (23.859) provide the data path connectivity compressed till local network, although the signaling for that remain intact like a normal PDN connectivity. This could be taken in principle for the convergence of two networks at local network level.

We believe strongly that Service provider can create yielding business case around a associated Wi-Fi network. The convergence at RAN and at network layer level would be a pure software solution with existing infrastructure and can be a ready to move solution. Other specific approaches will be requiring necessary ecosystem around for their success.

We are actively seeking the support and sponsors to extend our POC work. We need support from service provider and OEMs for widen-out our POC development. Please feel free to contact us for detail and discussion.

Note : We are strongly pursuing on our believe on next gen network with a central theme of "Homogeneous connectivity through heterogeneous networks" the solution for Wi-Fi convergence to cellular is inclined to this theme. We believe technology like MEC is going to give a master boost to our center theme.

Next Generation Mobile Networks Evolution with Convergence, Cloud and SDN

TAGS: NGN, HetNet, Cloud, SDN, smallcell, WiFi, offload, mobile core, mobile broadband, wireless networks

Humongous flood of smart devices and next generation applications and also the reach of networking to small piece of appliances the demand of the data traffic is being increased multi-fold on service networks. Networks are no longer remained and will remain an entity of just to provide connectivity but also to understand the whole eco system and have to evolve at every level.

These eco system requirements are coming up with new approaches, generating new challenges and enigmas. To cater these eco system requirements the concept of heterogeneous network at access and unified core network could be evaluated. Where there will be convergence at the access for heterogeneous radios and unified core for the unification of policy, provisioning and mobility.

As devices are already there or coming up with all the new interfaces along with 2G/3G or LTE and WiFi with hotspot2.0 -- Passpoint etc. These technologies are being utilized to provide heterogeneous network connectivity and to be managed to cater the various race and cast of traffic that the devices generate, depending on the services and connectivity.

This will not only provide the capacity enhancement but also efficiency of individual access network connectivity. It will result in total convergence at radio access and there would be selection mechanism to decide the network at access for specific service and related traffic forwarding to and from core network which is providing public network connectivity.

The complexities which will arise on such network evolution would be for the demand for flexibility, elasticity, unification, ubiquitous and seamlessness. The feasibility of addressing such requirements lies in the paradigm shift of complete separation of control and data plane and in such endeavour cloud and SDN are going to provide encouraging enabler. In such endeavour the control plane would fall to cloud and data plane would be on SDN.

Coming to more specifics on network architecture, Access stratum and non-access stratum (core) paradigm is going to hold with such enhancements. As access stratum would be more about access selection and capability associations etc. whereas core would be more on service delivery, policy, mobility and network connectivity.

on mobility front also shift will happen from hierarchical mobility to network mobility and hierarchical mobility will stand with the best limited reach at access stratum. That means to say that IETF protocols may dominates in the whole mobility aspects of the systems.

While looking at the upcoming thrust and compulsion for evolution in network architecture, a solution approach for the same could be contemplated. This is briefly defined as convergence at access stratum, and non-access stratum with unified core where an  ‘Access Controller Agent’ will help to select the access network and also will construct the core by finding the required association of cloud and SDN. The figure below depicts the emerging mechanism.

 The separation of control plane and data plane in next generation mobile core network architecture has already happened to some extent in EPS or SAE. Thanks to All IP paradigm in mobile networks to make this happing. Coming of cloud and SDN is going to corroborate this paradigm shift to further extent.

As depicted in the Figure the access will be approaching to unified core through globally unique network service access identifier (NSAPI), which will be helping to find the right Access controller agent to find association between cloud for control plane and SDN realm for data plane. Agent is a variant of MME as most of the MME functionality will move to cloud it would be there with global attributes and domain specific scope to land to right cloud and enter into right SDN realm or domain. The networks of agents would provide the unification of policy and provisioning and mobility.

This paradigm shift which I refer to as “homogenous connectivity of heterogeneous networks” could be envisioned to transform the next generation network architecture to suit the emerging eco system.