Traditionally, operators provide HGWs to their customers to enable them to access their multiple services. Although this assures that the customer has access to a basic set of features, it imposes a huge cost on the operator due to the large number of equipments required. In addition, very often, HGWs have problems and require the operator to send a technician for HGW repair or replacement, which is very expensive. Furthermore, anytime operators need to launch new services, e.g. VoIP, IPTV or IPv6, this usually requires a hardware replacement, increasing the burden to create new services, considering both the cost and the operational agility perspectives.

In this context, the new vHGW paradigm intends to improve the overall picture, reducing costs and increasing the agility of service creation. In this model, a basic L2 pHGW (physical HGW) is deployed in the customer' home to provide connectivity, while a vHGW (virtual HGW) is moved to the cloud, implementing the traditional functions. In order to increase flexibility, multiple function are available as part of the vHGW, e.g. firewall, parental control or NAT, which can be orchestrated (chained) in a way that different customers/devices may have multiple set of functions and follow different sequences. In a model like this, the addition of a new feature is very straightforward and simple, not requiring any change on pHGWs at customers' homes.

The ongoing work is focusing on traditional HGW functions. However, this model can be extended to other functions, providing a more customized and agile service environment. This can also be extended to mobile and corporate segments, to a large number of OTT services, or to network control functions, creating new business models based on function chaining and per customer/device/user customization. This is just the beginning of a new era of service agility and customization.

5G research has just started; the 5G requirements have been defined and there is already a vision of the 5G network. Moreover the key concepts and technologies that will be used for 5G have been already identified. The important role of SDN, SDR, NFV, virtualization an clouds in 5G is already known. In this talk the impact of software based solution on 5G networks will be discussed. There will be presented what these new approaches will bring to 5G and how they fundamentally change the system design in comparison to 4G and previous generations of mobile systems. Some design dilemmas will be presented as well as benefits and caveats  related to software implementations of complex, high speed and highly reliable real-time distributed systems (e.g. 5G).  There will be also presented how the complexity of software in 5G can be addressed and what can be the role of advanced and intelligent data processing using autonomic/cognitive and big data techniques. The key problems related to the design of Soft5G will be listed. Moreover  the impact of software on 5G standardization and business model will be also presented. Finally it will be explained why 5G-like networks can be deployed before 2020 and why 5G will probably be … the last generation of mobile networks.