Introduction to Evolved Packet Core

This white paper provides a brief introduction to Evolved Packet Core — a new mobile core for LTE (News - Alert). Herein, key concepts and functional elements-EPC gateways (Serving Gateway and Public Data Network Gateway), Mobility Management Entity (MME) and Policy and Charging Rules Function (PCRF)-are outlined, as well as key changes in LTE requirements imposed on the Evolved Packet Core, and the deployment challenges.

EPC is a new, all-IP mobile core network for the LTE — a converged framework for packet-based real-time and non-real-time services. It is specified by 3GPP Release 8 standards (which have been finalized in Q1 2009).

The EPC provides mobile core functionality that, in previous mobile generations (2G, 3G), has been realized through two separate sub-domains: circuit-switched (CS) for voice and packet-switched (PS) for data. As shown in Figure 1, in LTE, these two distinct mobile core sub-domains, used for separate processing and switching of mobile voice and data, are unified as a single IP domain. LTE will be end-to-end all-IP: from mobile handsets and other terminal devices with embedded IP capabilities, over IP-based Evolved NodeBs (LTE base stations), across the EPC and throughout the application domain (IMS and non-IMS).

EPC is essential for end-to-end IP service delivery across LTE. As well, it is instrumental in allowing the introduction of new business models, such as partnering/revenue sharing with third-party content and application providers. EPC promotes the introduction of new innovative services and the enablement of new applications.

Click here to enlarge Figure 1

 

The EPC is a new, high-performance, high-capacity all-IP core network for LTE. EPC addresses LTE requirements to provide advanced real-time and media-rich services with enhanced Quality of Experience (QoE). EPC improves network performance by the separation of control and data planes and through a flattened IP architecture, which reduces the hierarchy between mobile data elements (for example, data connections from eNodeB only traverse through EPC gateways). Figure 2 shows the EPC as a core part of the all-IP environment of LTE.

Click here to enlarge Figure 2

 

The introduction of the EPC and all-IP network architecture in the mobile network has profound

implications on:

• Mobile services, as all voice, data and video communications are built on the IP protocol

• Interworking of the new mobile architecture with previous mobile generations (2G/3G)

• Scalability required by each of the core elements to address dramatic increases in number of direct connections to user terminals, orders of magnitude of bandwidth increase, and dynamic terminal mobility

• Reliability and availability delivered by each element to ensure service continuity

To address a radically different set of network and service requirements, the EPC must represent a departure from existing mobile networking paradigms.

Introduction of EPC with LTE in many ways represents a radical departure from previous mobile paradigms:

• End of circuit-switched voice: LTE uses a new paradigm for voice traffic — VoIP. This ends a period of more than 20 years during which one application dictated the whole network architecture. EPC treats voice as it is: just one of many IP-based network applications, albeit an important one that requires superb packet network performance and one that is responsible for significant operator revenues.

• Evolved wireless broadband: LTE must match and exceed the QoE of wireline broadband. This is quite different from providing best-effort low-speed web browsing or Short Message Service (SMS) — two data applications for which the existing PS mobile cores were optimized.

• Mobility as a part of the core network: In LTE, all mobility management is moved into the mobile core and becomes the responsibility of the MME. This is a consequence of the split of functions previously performed by the RNC and NodeB/BTS. The MME requires a control plane capacity that is an order of magnitude larger than the SGSN or PDSN, and must ensure interworking with 2G/3G legacy mobile systems.

• End-to-end QoS becomes essential: LTE must provide superior end-to-end QoS management and enforcement in order to deliver new media-rich, low-latency and real-time services. There is an expected move from four classes of service (CoS) available in 3G to nine QoS profiles with strict performance targets. This must be achieved while ensuring scalability of users, services and data sessions. In addition, although not a part of the 3GPP Release 8 specification set, DPI and other advanced packet processing are required at the beginning.

• Policy management and enforcement: In LTE, service control is provided via the Policy and Charging Rules Function (PCRF). This is a big change from previous mobile systems, where service control was realized primarily through UE authentication by the network. PCRF dynamically controls and manages all data sessions and provides appropriate interfaces towards charging and billing systems, as well as enables new business models.

LTE requires significantly more capacity in both the data plane and control plane. The existing 2G/3G mobile core elements cannot fully address LTE requirements without a series of upgrades to the platforms. Most of the existing platforms that are envisaged to evolve to address LTE requirements are ill-suited for high-capacity packet processing. Scaling the packet processing requirements on these platforms results in higher consumption of system capacity, high latency, low performance and dramatic performance/feature tradeoffs. In some cases, performance drops more than 50% or more when features like charging are enabled. Legacy core platforms must dramatically change their product architectures to handle LTE, and even with these architectural changes, they are only a stop-gap solution that may require complex upgrade scenarios to address LTE scalability and performance requirements.

Alcatel-Lucent (News - Alert) believes that delivering a new LTE mobile core in the form of newly deployed, purpose-built EPC elements is essential to ensure superior network performance and quality of services and to minimize overall business costs. At the same time, this forward-looking EPC is a cornerstone for further business evolution in mobile networks, allowing the creation of new business models and facilitating rapid deployment of new innovative services.

The EPC is realized through four new elements:

• Serving Gateway (News - Alert) (SG-W)

• Packet Data Network (PDN) Gateway (PGW)

• Mobility Management Entity (MME)

• Policy and Charging Rules Function (PCRF)

While SGW, PGW and MME are introduced in 3GPP Release 8, PCRF was introduced in 3GPP Release 7. Until now, the architectures using PCRF have not been widely adopted. The PCRF’s interoperation with the EPC gateways and the MME is mandatory in Release 8 and essential for the operation of the LTE.

The SGW is a data plane element whose primary function is to manage user-plane mobility and act as a demarcation point between the RAN and core networks. SGW maintains data paths between eNodeBs and the PDN Gateway (PGW). From a functional perspective, the SGW is the termination point of the packet data network interface towards E-UTRAN. When terminals move across areas served by eNodeB elements in E-UTRAN, the SGW serves as a local mobility anchor. This means that packets are routed through this point for intra E-UTRAN mobility and mobility with other 3GPP technologies, such as 2G/GSM and 3G/UMTS. Figure 3 shows the Serving Gateway.

Like the SGW, the Packet Data Network Gateway (PDN GW) is the termination point of the packet data interface towards the Packet Data Network(s). As an anchor point for sessions towards the external Packet Data Networks, the PDN GW supports:

• Policy enforcement features (applies operator-defined rules for resource allocation and usage)

• Packet filtering (for example, deep packet inspection for application type detection)

• Charging support (for example, per-URL charging)

In LTE, data plane traffic is carried over virtual connections called service data flows (SDFs). SDFs, in turn, are carried over bearers — virtual containers with unique QoS characteristics. Figure 4 illustrates the scenario where one or more SDFs are aggregated and carried over one bearer.

Click here to enlarge Figure 4

 

One bearer, a datapath between a UE and a PDN, has three segments:

• Radio bearer between UE and eNodeB

• Data bearer between eNodeB and SGW (S1 bearer)

• Data bearer between SGW and PGW (S5 bearer)

Figure 5 illustrates three segments that constitute an end-to-end bearer. The primary role of a PGW is QoS enforcement for each of these SDFs, while SGW focuses on dynamic management of bearers.

Click here to enlarge Figure 5

 

The Mobility Management Entity (MME) is a nodal element within the LTE EPC. It performs the signaling and control functions to manage the User Equipment (UE) access to network connections, the assignment of network resources, and the management of the mobility states to support tracking, paging, roaming and handovers. MME controls all control plane functions related to subscriber and session management.

MME manages thousands of eNodeB elements, which is one of the key differences from requirements previously seen in 2G/3G (on RNC/SGSN platforms). The MME is the key element for gateway selection within the EPC (Serving and PDN). It also performs signaling and selection of legacy gateways for handovers for other 2G/3G networks. The MME also performs the bearer management control functions to establish the bearer paths that the UE/ATs use.

The MME supports the following functions:

• Security procedures: End-user authentication as well as initiation and negotiation of ciphering and integrity protection algorithms.

• Terminal-to-network session handling: All the signaling procedures used to set up packet data context and negotiate associated parameters like QoS.

• Idle terminal location management: The tracking area update process used to enable the network to join terminals for incoming sessions.

The major improvement provided in Release 7 of 3GPP in terms of policy and charging is the definition of a new converged architecture to allow the optimization of interactions between the Policy and Rules functions. The R7 evolution involves a new network node, Policy and Charging Rules Function (PCRF), which is a concatenation of Policy Decision Function (PDF) and Charging Rules Function (CRF).

Release 8 further enhances PCRF functionality by widening the scope of the Policy and Charging Control (PCC) framework to facilitate non-3GPP access to the network (for example, WiFi (News - Alert) or fixed IP broadband access).

In the generic policy and charging control 3GPP model, the Policy and Charging Enforcement Function (PCEF) is the generic name for the functional entity that supports service data flow detection, policy enforcement and flow-based charging. The Application Function (AF) here represents the network element that supports applications that require dynamic policy and/or charging control. In the IMS model, the AF is implemented by the Proxy Call charging Session Control Function (P-CSCF).

Figure 6 shows how PCRF interfaces with other EPC elements.

As EPC radically changes key networking paradigms for previous generations of mobile core networks, the introduction of Evolved Packet Core must successfully address a number of technological challenges.

In LTE, there are significant technological advances on the radio side (eNodeB). LTE provides more efficient use of the spectrum with wider spectral bands reserved for LTE. This results in greater system capacity and performance. At the same time, the mobile core needs to change to provide higher throughput and low latency; both should come as results of the simplified and improved flat all-IP network architecture.

Delivery of the superior LTE solution and the introduction of new technologies — both on the radio side and in the core — is an important task. The existing (2G/3G) mobile cores, designed and engineered for low-speed, best-effort data, cannot provide the required scalability or ensure high performance.

EPC needs to address a number of key IP aspects for deployment in LTE:

• IP routing and network addressing, and real-time management if IP domains are large

• Centralized vs. distributed network architecture (for MME, SGW and PGW deployment)

• IPv6 (strategy, introduction, coordination with IPv4)

• End-to-end QoS deployment and coordination with underlying transport

• L2 vs. L3 connectivity at the transport layer (eNodeB, SGW, PGW, MME)

• End-to-end security for data and control plane

• Interconnectivity to external networks and VPNs

• Deep Packet Inspection, Lawful Interception

There is a stringent set of requirements for reliable, scalable, high-performance elements due to the dynamic nature of user mobility in LTE, coupled with the short duration of multiple data sessions per UE and large-scale deployment targets. To meet these demands, the EPC elements must be best in class, with superior IP performance. On the product (network element) level, in order to address all these important aspects of EPC’s introduction, a new generation of purpose-built, highly scalable mobile core equipment is required, and a strong IP expertise is needed.

Putting all these elements together is as important as delivering all elements with the needed carrier-grade features for LTE. EPC elements must interwork in full harmony as the fairly complex network procedures involve all EPC elements, in both control and user planes. The EPC must address the demanding requirements for multi-dimensional, dynamic management of mobility, policies and data bearers — in an orchestrated manner that enables the highest LTE performance, while ensuring interworking and interoperability with legacy 3G/2G systems. Figure 7 illustrates the dynamic nature of policy and mobility management in LTE.

Copyright © 2009 Alcatel-Lucent. All rights reserved.