YateUCN, the EPC cloud

The YateUCN is an LTE EPC that unifies all the functions of a conventional LTE core network into a single server. A single YateUCN unit combines the MME, SGW, PGW, PCEF and PCRF functions. A pool of YateUCN servers provide seamless horizontal redundancy, scalability and load balancing in the LTE core network. The YateUCN also replaces the latest core network approaches to design and management, such as network function virtualization (NFV) that virtualizes the functions of the network’s nodes in software or software-defined networking (SDN) that splits the control plane from the user plane.

A conventional LTE core network has many components and each requires a back-up node for redundancy. To ensure load balancing, operators need to deploy load balancers and external servers, which only add to the complexity count.

The YateUCN servers that form the EPC cloud have a many-to-many relationship with the eNodeBs, are equal at application lever, and eliminate the single point of failure between the RAN and the core network. All these characteristics allow for horizontal redundancy, load balancing, easy management and an overall simplicity within the network.

yucn_redund_epc_2015-11-18_version1.2_compare1.1

Redundant EPC

In conventional LTE networks, all core network components (MME, SGW, PGW, PCEF, PCRF) need to be duplicated to ensure redundancy and synchronization in case of failure. With a YateUCN-based core network, operators add extra servers to the existing pool to increase the network’s overall capacity.

To achieve redundancy, subscribers get assigned to random servers from the YateUCN pool. If a YateUCN fails, all the devices served by that unit are automatically re-assigned to other available YateUCN servers, as seen in the diagram.yucn_redund_epc_2015-11-18_version1.3

LTE core network cloud

A cluster of YateUCN servers act as an LTE core network cloud, providing all the EPC services: mobility, authentication, quality of service, routing upload and download IP packets, IP address allocation and more. Mobile operators eliminate the occurrence of a single point of failure between the RAN and the core network because YateUCN servers are equal at application level, and have a many-to-many relationship with the eNodeBs.

By removing the single point of failure possibility, mobile carriers can build scalable and considerably leaner core networks, while also providing load balancing for enhanced flexibility.

Each YateUCN core network server is implemented in off-the-shelf servers commodity software (Linux), offering a shorter lead time, more servicing options and faster replacement time.

Integration in existing LTE networks

The entire LTE EPC layer is implemented in a single YateUCN unit, meaning that it replaces the MME, the SGW, the PGW, the PCEF and the PCRF units of conventional networks.

The YateUCN is compatible with any generic LTE RAN and core network component. It supports the S1-AP and GTP interfaces between its MME and SGW functions and the eNodeBs. The YateUCN uses Diameter (S6a) to connect to an existing generic Home Subscriber Server (HSS). The unified server can connect to external PGW and SGW via the S5/S8 interface. To link to an existing IMS, to the Internet or to an MMS service, the YateUCN uses the SGi interface. Finally, to interrogate an external Equipment Identity Register (EIR) about blacklisted IMEIs, the YateUCN uses S13 over Diameter.

Additionally, the YateUCN also implements the IMS functions necessary for deploying VoLTE, but this will be detailed in a future article. In the meantime, previous blog posts have detailed the YateUCN’s VoLTE call with an iPhone 6 or how the YateUCN handles SRVCC.

A few final words

A unified core network server that delivers redundancy, scalability, load balancing and flexibility allows mobile operators to tap new core network equipment innovations and reduce their CAPEX. Easily integrated in an existing LTE network, the YateUCN makes it possible for mobile carriers to optimize their networks and replace solutions typically characteristic to the recent mobile LTE deployments, such as NFV or SDN.

YateUCN – the redundant MSC/VLR

Traditionally, the redundancy of the Mobile Switching Center / Visitor Location Register (MSC/VLR) is obtained through redundant dedicated hardware and software. The problem lies in the Abis + A interfaces (BSSAP protocol) which do not allow a base station to move easily to another MSC/VLR. To overcome this problem, we decided to use the SIP protocol (with enhanced features for GSM operations), which allows a YateBTS SatSite base station to move to another YateUCN (MSC/VLR) automatically and quickly.

Redundant GSM MSC/VLR

In a GSM network deployed with YateBTS-based SatSite base stations and YateUCN core network servers, each subscriber is randomly assigned to a YateUCN from a pool of core network servers. Operators can increase the redundancy of the pool by simply adding additional YateUCN servers for excess capacity in case a unit malfunctions. Therefore, handsets connected to a single SatSite can be served by multiple identical YateUCNs, while, at the same time, a single YateUCN serves multiple SatSite units. If a YateUCN server fails, all the mobile devices served by it are automatically moved to the other available YateUCN servers. When the subscriber is registered to the new YateUCN, its location is updated in the HLR.

In short, YateUCN and the GSM YateBTS SatSite base stations form a many-to-many relationship, and this was made possible though a number of characteristics.

  • The GSM YateBTS SatSite implements all the functions of a conventional Base Station Controller (BSC).
  • The A interface (between the BSC and the MSC/VLR) was replaced with SIP (between YateBTS and YateUCN), making it possible to quickly re-associate handsets with different YateUCN servers.
  • All YateUCN servers are identical units that support many core network functions. Operators will only have to duplicate one component, as opposed to multiple in conventional networks.
  • YateUCN is implemented in commodity hardware (off-the-shelf servers) and software (Linux), delivering a shorter lead time, more servicing options and faster replacement time.

The diagram below illustrates the technology perfectly.

yucn_msc-vlr_2015-11-4_version1.1

Integration in an LTE core network

The SatSite and the YateUCN components are easy to integrate into existing LTE networks because 2G services are delivered in SIP. These GSM services are integrated into a 4G network by employing the same GTP and IMS interfaces that are typically used in conventional EPC/IMS core networks.

YateUCN implements the same SIP switch to both provide GSM services and connect to an existing IMS network. Thus, with both the SIP-powered RAN and core network products, operators’ migration from GSM to LTE turns into a much simpler process.

Integration in a GSM core network

As an MSC/VLR, YateUCN performs all the functions of other MSC/VLRs: mobility, authentication, speech call and SMS routing. The server supports authentication of handsets using the SIM/USIM (EAP-SIM/EAP-AKA) and SIP AKAv1-MD5 algorithms.

YateUCN can connect to any standard HLR via the SS7 MAP protocol, and to other MSCs and GMSCs through the MAP-E protocol, allowing it to be in any conventional GSM network.

A few final words

YateUCN brings something new to GSM equipment: affordable redundancy, scalability, and uncomplicated management into a single core network server. It is easy to integrate to existing GSM and LTE networks and can be easily upgraded with new features within the same hardware and allows operators a seamless network extension.

Rethinking redundancy: a new approach to core networks

Mobile communications must provide uninterrupted mobile service at all times, but the costs to create network redundancy with current conventional equipment are restrictive. YateUCN unified core is a profitable and flexible solution for redundancy in 2G and 4G mobile networks.

Network redundancy ensures that as technology advances, the capacity of network infrastructures to support more subscribers without blackouts adapts accordingly. YateUCN is a unified core network allowing resiliency in 2G/LTE mobile networks using YateBTS and YateENB SatSite. SatSite acts as a BTS/BSC communicating directly with the MSC/VLR/SGSN/GGSN and EPC in YateUCN.

As a software-defined core solution, YateUCN replaces the heavy, expensive core equipment used in conventional networks with smaller, affordable, and easy-to-manage equipment. It is a software implementation of 2G and LTE core network layers, operating on commodity hardware.

yucn_redund_2015-7-16_draft4.2_pic2

In typical networks, redundancy is achieved by supplying an additional core server for any given core server, causing costs to more than double, since supplementary costs for the configuration of back-up servers add up to the capital expenses.

YateUCN implements the core network functions and protocols in software, enabling any other YateUCN node to take over extra-traffic in case of failure of a node, or if the network capacity needs to be increased.

While conventional MSC/VLR in data centers are limited to serving a given number of BSCs in a defined geographical area, in a YateUCN – SatSite network the base station allows a device to connect to any YateUCN node in the network, irrespectively of the geographical location of the device/BTS and of whether the network is 2G or 4G. A list of available YateUCN units is configured in each YateBTS/YateENB SatSite.

Core equipment is usually designed to allocate specific core network functions (authentication, mobility, call setup, data routing) to separate nodes. Such equipment is heavy due to the large number of components, increases lead time, and requires separate back-up equipment for each node.

YateUCN unifies both GSM and LTE core layers, meaning that a single alternate YateUCN server provides full redundancy for any other server in the network. If a failover should occur in a YateUCN node, a device can register to a different YateUCN, remaining attached to the same base station, as shown below.

yucn_redund_2015-7-15_draft4.1

A new YateUCN is chosen from the list of YateUCN units held in the base station. If a mobile device remains connected to the same BTS, registration to the MSC/VLR in the new YateUCN is performed whenever the device communicates with the network to perform an action. Registration to the new YateUCN is updated in the HSS/HLR.

If the device roams to an area served by a different BTS, they will connect to the new SatSite, but will remain connected to the YateUCN currently serving it, and a new query in the HLR is not required. This reduces the load on the HLR and allows it to support a higher subscriber capacity. This can be seen below:

MS connecting to a new YateUCN

MS connecting to a new YateUCN

Increasing traffic to a YateUCN core server is easily performed because YateUCN communicates with 2G base stations using SIP and GTP, and with eNodeBs over SIP/S1AP/GTP. SIP and GTP signalling protocols have the advantage of scalability and interoperability, allowing different service requirements to be served at the same time and with the same quality standards.

Because YateUCN uses commodity hardware, operation and servicing can be managed remotely, with minimal external support, significantly driving operational costs down. YateUCN provides simplicity and cost-effectiveness to building redundancy in mobile networks so that operators can provide high-quality service at all times.

Revolutionary Features in YateBTS 4

YateBTS 4 introduces redundancy support for YateUCN. This is a revolutionary feature that deserves some explanation.
redundancy_scheme_1.2
Redundancy failover and load balancing.
A YateBTS network is a many-to-many mapping between YateBTS base stations and YateUCN core network servers. Core network loads are distributed evenly and are redistributed automatically if a server fails. The resulting network is resilient and easy to manage. To add more capacity, just add more servers. If a server fails, its users are shifted to other units within a few seconds. This is a sharp contrast to the conventional mobile network, a hand-configured, tree-like hierarchy, with points of failure that become more concentrated as you move toward the HLR.
Release 4 also introduces handover support, which is revolutionary in its own way because it is done very differently from conventional GSM. Because YateBTS has no BSCs, handover is a peer-to-peer operation. The result is a 2.5G radio network that behaves much more like LTE, which is one of the innovations that make the Unified Core Network possible.
 handover_scheme_2.2Handover support is done using SIP
YateBTS Release 4 has been available since July, but got little fanfare at the time. That won’t be the case with Release 5, which will be ready soon, so watch this space.