WiFi/WiMAX Heterogeneous Seamless Handover

At the Mobile World Congress which begins today in Barcelona, Intel will be showing a demo of our research to perform a heterogeneous seamless handover between a WiFi and WiMAX network. In this blog I will describe our research work to improve and optimize seamless transition across these networks which was done in close collaboration with Nokia R&D and Nokia Siemens Networks R&D.

In this effort Nokia Siemens Networks focused on the network infrastructure providing the Information Service and Command Service while Intel and Nokia focused on the mobile client. Intel provided Layer 2.5 functionality including link layer events, client information and command service, and Nokia provided Mobility Management. The demo gives a first glimpse to a possible future solution, and we’ll continue the studies and experiments to show the measurable improvements provided by the technology. Watch the video below to see researcher Vijay Kesavan demonstrating a WiFi/WiMAX handover. You can see yours truly at the other end of Vijay’s video call.

Innovations in small form factor multi-radio consumer devices (Smartphones, Mobile Internet Devices, UMPCs, etc) are fueling the end-user appetite for Anytime Anywhere access to services. ABI Research estimates the market size for MIDs (having WiFi and WiMAX/3G) to grow from 3.4 Million Units in 2008 to over 89 Million Units in 2012. The demand for ubiquitous services over emerging multi-access radio networks can be addressed via seamless handovers so consumers don’t suffer service disruptions when roaming and running multimedia applications.

For the mobile device the steps to perform a horizontal (within a same network type) or vertical (across heterogeneous network) handover can be classified into three logical tasks shown in Figure 1 which in chronological order are, i) “when and why” should the device transition to a new network, ii) “where” should the device transition to, and iii) “how” should the device transition between networks to maintain connectivity and session persistence. The “when and why” is triggered when the mobile device receives an indication that a transition should take place. These can include external conditions like signal degradation or network congestion, discovery of a more suitable network providing lower cost, higher bandwidth or better energy efficiency, or be user or network initiated. The “where” step is when the mobile device selects the next network to connect to, and the “how” defines how the device performs the transition, e.g. either doing a horizontal or vertical handover.

The network can facilitate the mobile device network selection by providing a network map describing the presence and characteristic of networks available in the vicinity of the mobile device. Additionally the network itself can initiate a handover request when for example it detects that the load on one of the access networks is nearing capacity. Watch the video below to hear researcher Vijay Kesavan discussing the technology behind a WiFi/WiMAX handover.

“When and why?” to handover. The goal is to provide accurate and predictive link layer information about an impending degradation or loss of signal to start a handover – for example a make-before-break operation. A three step process predicts with a high level of accuracy the signal quality direction. We (i) smooth out the signal characteristics, (ii) predict the signal level in the near future and (iii) perform a trend analysis, focusing on methods that provide accuracy with a low computation overhead. We use exponential average for smoothing, “Straight Line” for prediction and Fast Fourier Transform for trend analysis on both the short and long term window. Results show that we predict link layer degradation on WiFi and WiMAX with about 90% accuracy and provide this indication about 800 msec to 1 second before the loss happens thereby providing a significant lead time to take action.

<System Architecture
Figure 1: System Architecture

“Where?” to handover to? Once indication of an impending disconnection is received, the next decision is to select another suitable network to connect to. In the wireless environment the ability to discover networks depends on the detection and proximity of the signal source and the environment, e.g. interferers, obstacles, etc. Mobile devices periodically scan to “search” wireless networks on each of their interfaces thus consuming power. This is where the information learned from the network on the existing connected channel, e.g. a network information map describing the presence and characteristics of other networks, helps the mobile device improve its energy efficiency by scanning only on appropriate interfaces at specific geographic locations. To evaluate the best network to go to, a cost function taking into account network level (e.g. bandwidth, delay, security, etc.), platform level (e.g. battery load, thermal, etc.), user’s preferences (e.g. monetary cost, operator, etc.) and network preferences defined by network operator/virtual network operator/enterprise is used.

“How?” to handover to? This is the execution part of the handover process. There are many methods available for handling session handover, e.g. Mobile IP (MIP), Proxy Mobile IP, Session Initiation Protocol (SIP), Voice Call Continuity (3GPP VCC). We used a SIP based A/V application.

Fig2_Handovers.jpg
Figure 2: Heterogeneous Handover with predictive triggers timing diagram

Using the optimizations mentioned earlier we can see in Figure-2 that we can start a make-before-break where the two radios are concurrently active with enough lead time to support seamless handover for multimedia sessions. (It should be noted that authentication has not been taken into account here). Our implementation is based on the IEEE 802.21 Draft specification for both client and network components (Information Server and Command Server), but this is only one option among many to achieve this.

Christian Maciocco

Research Scientist, Communications Technology Lab

Christian joined Intel in 1994 and has held positions in research, architecture development, and engineering management. He has worked on a number of emerging technologies at Intel including IP Telephony and media streaming over the Internet, HDTV / DTV data broadcasting, optical networking, wireless communication and most recently Energy Efficient Communication. He has represented Intel in various standards and industry organizations, including ATM Forum, ATSC, IETF, OIF, and SDR Forum. He has 5 patents granted, 28 pending and co-authored multiple publications. He earned a Diplôme d’Ingénieur – Ecole Speciale de Mecanique & Electricité, SUDRIA – Paris

10 Responses to WiFi/WiMAX Heterogeneous Seamless Handover

  1. Mika Forssell says:

    Topic is highly relevant as networks will be increasingly heterogeneous with multiple network technologies and players involved. Despite of heterogeneous networks, end users want to access their services anytime and anywhere, and therefore we need to do the same for Internet and (broadband) data services as cellular did for voice services. Thus networks of different technologies need to be bridged for seamless access to services provided by multiple service providers.
    We can’t leave the connectivity complexity for the end user to handle as is the case in today’s networks and devices, where user has to tell the device of how to get connected. Complexity needs to be hidden from end user via automated connection operations and this demo seems to do just that via preferences and network selection intelligence.
    Also making efficient use of available network and device resources is highly important. When traffic volumes increase, distributing traffic across available networks makes sense to avoid congestion – and this requires network involvement. Also if device battery consumption can be decreased through use of new technologies, this allows taking real advantage of strengths of different networks.
    Demo seems to provide nice package as service continuity is also addressed in the demo. If connections would break when moving between two networks, mobility benefits were compromised.

  2. Jim Callaway says:

    Simply put – without the realization that this is neccessary and the solution therein – we are not going anywhere near alternative service provider national mobility.

  3. Janne Tervonen says:

    Seamless service continuity in multi-access (or heterogeneous) radio networks is a hot topic. In order to have a great user experience when changing from a network to another, we need the support from both the device and network. I hope we can see this in products in the near future!

  4. Ganesh says:

    The demo was interesting, as we could see the seamless handover of video over heterogeneous networks. If this scenario could be tested in real time with the help of a service provider, it will further help to improve the overall handover scenario and increase its efficiency.

  5. Rute Sofia says:

    It is good to finally see this kind of initiative, even if in a very controlled environment.
    However, a remark: basing the global handover upon a SIP solution to provide seamless handovers will work as long as no TCP traffic is involved, and only for SIP-based applications…which means that such solution wouldn’t work for *any* A/V service available today in the Internet…a bold move would really to try to use technology (already available today…) that could assist seamless handovers independently of application. So that it could support current *and* future applications.

  6. A. Althagafi says:

    Handover certainly is today’s hot topic, regardless of the network access user uses. Knowing the advent of LTE, 4G and Wimax, and the strong presence of WiFi, it becomes imminent to bridge between such technologies to let users enjoy 4-play applications seamlessly

  7. William Halverson says:

    There are a lot of places in developed countries that still don’t have broadband internet access. I own property in upstate New York where there is no DSL or CATV broadband … and there are no plans to ever put it in.
    This would be a great solution for ‘co-op’ networks in mountainous, sparsely settled areas.

  8. pest control las vegas says:

    multiple network technologies and players involved. Despite of heterogeneous networks, end users want to access their services anytime and anywhere, and therefore we need to do the same for Internet and (broadband) data services as cellular did for voice services. Thus networks of different technologies need to be bridged for seamless access to services provided by multiple service providers.
    We can’t leave the connectivity complexity for the end user to handle as is the case in today’s networks and devices, where user has to tell the device of how to get connected. Complexity needs to be hidden from end user via automated connection operations and this demo seems to do just that via preferences and network selection intelligence.
    Also making efficient use of available network and device resources is highly important. When traffic volumes increase, distributing traffic across available networks makes sense to avoid congestion – and this requires network involvement. Also if device battery consumption can be decreased through use of new technologies, this allows taking real advantage of strengths of different networks.
    Demo seems to provide nice package as service continuity is also addressed in the demo. If connections would break when moving between two networks, mobility benefits were compromised.

  9. Gopal says:

    Thanks for the interesting article and demo. While I can see the process of handover from Wifi -> Wimax, I was wondering if this handover would work seamlessly from wifi -> wifi (considering that wifi adaptors in the Mobile unit are programmed to do break-before-make). I suspect that currently this demo setup would not be able to achieve that (would be glad to be corrected and provided pointers on how do you go about achieving it).
    My understanding so far is that handover across AP’s in a wifi domain (even the same subnet) takes around 5-10 seconds. If an application (such as playing an online game on the MU) does not support sessions based reconnections, then “seamless” handover might be a problem (i.e., the online gamer will have restart the game and reconnect to the gaming server).
    There is a similar project called “OpenRoads” (Stanford University) which demoed wifi-wimax handover. Was wondering if this similar to that or radically different in any way.