ITRM20130281A1 - METHOD AND APPARATUS FOR THE SEAMLESS HANDOVER "SIDE NETWORK", INVISIBLE "TERMINAL SIDE", FOR TERMINALS COMPLIANT WITH THE IEEE 802.11 STANDARD, BASED ON A LOGIC OF CONTROL OF THE DISTRIBUTED AND SYNCHRONIZED HANDOVER, INDEPENDENTLY FROM THE DORSAL TYPE - Google Patents

Description

DESCRIPTION

Description of the invention entitled:

"Method and Apparatus for Seamless Handover" network side ", invisible" terminal side ", for terminals compliant with the IEEE 802.11 standard, based on a distributed and synchronized handover control logic, regardless of the type of network backbone"

FIELD OF THE INVENTION

The present invention relates in general to communication networks based on any IEEE 802.11 specification, such as but not limited to IEEE 802.11 a, b, g, n, ac and the like, also known as "WiFi" networks, and in particular refers to the transfer of PDU flow management from one Access Point (AP) to another, without interruption in PDU exchanges from and to each terminal compatible with the standard, better known and henceforth referred to as "seamless handover" .

STATE OF THE ART

Cellular networks achieve easy handover between radio base stations by sharing information about a given mobile device. This session data is used for routing and is updated every time a terminal changes cells [Bejerano et al. 2002; Chiasserini 2002]. On the other hand, the IEEE 802.11 standard lacks the mechanisms available in the protocols of modern cellular networks.

Mishra et al. Analyzed link layer handoff performance in IEEE 802.11 devices [Mishra et al. 2003]. About 90% of the handover delay is attributable to the scanning of the band that a radio device performs to find its nearby APs. Their experiments also show that the actual handover delay can vary greatly depending on the manufacturers used for the terminal network card and the AP. Vatn also studied the latency effects of a wireless handover on voice traffic [Vatn 2003]. The conclusions echo those of Mishra et al. about the fact that handover latency can vary greatly depending on the hardware vendor used. Since our approach does not require reassociation during handover, we do not suffer from these manufacturer-specific delays.

Ramani and Savage demonstrated that fast handover-level connection is possible in IEEE 802.11 networks when the terminal checks the signal quality of the APs and uses a fast scan mechanism to listen to all available APs to choose the best [Ramani and Savage 2005]. Their SyncScan system achieved an impressive handover time as low as 5ms. Scan speed is achieved through changes to the terminal's network card driver. In contrast, our approach uses any unmodified IEEE standard 802.11 terminal.

Two well-known general approaches for handover management in IP networks are IP Cellular [Valk 1999], and Hawaii [Ramjee et al. 1999]. In Hawaii, or Handoff-Aware Wireless Access Internet Infrastructure, messages are exchanged between the old gateway and the new gateway for packet forwarding.

IP Cellular establishes routes based on traffic from the terminal, and handover occurs when a cross-over router is reached. A comparison of these two protocols is presented in Campbell et al. [2002]. In a different approach to mobility proposed by Caceres and Padmanabhan, APs broadcast gratuitous ARP to their upstream router to create the illusion that mobile terminals are always connected to the wired network [Caceres and Padmanabhan 1998]. These approaches rely on the terminals to initiate the handover process, and are not concerned at all with addressing the handover delay present in IEEE 802.11 networks at the link level, when the terminals reassociate with another AP. Other approaches to handover, such as TMIP [Grilo et al. 2001; Yokota et al. 2002], and Sharma et al. [2004], improve handover latency in IEEE 802.11 networks, but do not exceed these limits. Other general approaches such as IDMP [Das et al. 2002], SMIP [Hsieh et al. 2003] and HMIP [Soliman and Bellier 2004] focus on a hierarchical approach to reduce the overall signaling load and improve scalability. Most of these approaches require software to be installed on mobile terminals. Instead, we teach an all "network side" method that is independent of the type of PDU exchanged with the mobile terminal and we propose a new approach to control the handover process directly from the network infrastructure without any interruption.

Seshan et al. [1996] used a multicast approach within the Dedalus project to ensure timely delivery of traffic to the terminal during a handover in an available cellular wireless network in 1996. Subsequently, Helmy et al. showed how a fast handover can be achieved in wireless networks by requiring mobile terminals to explicitly join a multicast group in which packets are tunneled through the infrastructure [Helmy et al. 2004]. Multicast during handover, called simulcast, is also used in S-MIP [Hsieh et al.

2003].

In at least one implementation, our approach does not require multicasting functions and supports terminals with any system architecture or operating system configured for the IEEE 802.11 standard "Infrastructure" operating mode (also known as BSS - Basic Service Set or AP mode). Although the examples presented refer to such a case, our method can also be applied to support terminals operating in the IEEE 802.11 standard in "ad-hoc" mode. The IEEE has also worked on standardizing handover for wireless IP networks at two different levels. The 802.11r standard aims to provide a fast transition between Basic Service Sets (BSS), allowing terminals to use their current access point as a gateway to other access points. The 802.21 standard instead aims to provide a handover between different types of networks, commonly known as “media-independent” or vertical handover. These approaches require changes to the 802.11 standard, and therefore for both access points and each end device. In our approach, no changes to the 802.11 standard are necessary, especially from the point of view of the mobile terminal.

Among the experimental test beds for mesh-type wireless networks, otherwise known as “mesh”, which support the mobility of terminals, we find MeshCluster [Ramachandran et al. 2005], and iMesh [Navda et al. 2005], both of which work with mobile terminals in infrastructure mode. Mesh-Cluster, which uses Mobile IP (MIP) [Perkins 1996], for intra-domain handover, shows a latency of about 700 ms for the delay due to reassociation to the access point and MIP registration. IMesh also offers handover using regular route updates or “Mobile IP”. Using ISO / OSI level 2 triggers for handover (terminal not moving), the handover latency in iMesh reaches 50-100 ms. The approach was then used in a more realistic environment for improving VoIP performance in mesh networks, with similar results [Ganguly et al. 2006]. Another mobility experiment is described in the Smesh project in [SMESH. 2010] for terminals that work only in "ad-hoc" mode.

We teach a method that can be used to create a seamless handover function capable of supporting real-time multimedia communications for IEEE 802.11 standard terminals, without the need to extend the IEEE 802.11 standard itself. In one embodiment, the method we teach provides fast link-level handover for IEEE 802.11 terminals working in infrastructure mode (BSS), controlling handover from the distributed network infrastructure of the APs and emulating a single large BSS network. , thus eliminating the need and perception of a handover from the point of view of the mobile terminal.

GENERAL DESCRIPTION

In a wireless communication network in which access points to the network, better known and henceforth referred to as "Access Points" (AP) exchange Protocol Data Units (PDUs) with devices such as, by way of example and not limited to, those adhering to the IEEE 802.11a / b / g / n / ac specifications, the teachings presented here describe a method of transferring PDU flow control from one AP to another while such a device moves from a source AP to a destination AP , which ensures the continuity of the flows themselves from the point of view of the terminal (the so-called seamless handover).

The IEEE 802.11 example is not limiting, since the teachings reported here are valid for any network that uses a probabilistic media access control protocol (MAC) on a shared transmission medium.

We call Base Station or “Base Station” (BS), a network node that, in one implementation, emulates the basic functions of an IEEE 802.11 AP and is configured to execute the control logic that implements the method we teach.

We call Mobile Station or "Mobile Station" (MS), a device which, in one embodiment and by way of example, acts as a terminal compliant with the IEEE 802.11 (STA) standard and moves between the radio coverage areas of the various BS, while continuing to behave as if it were always connected to the same AP.

We call the Cell Coverage Area (CCA) of a BS the area within which the signal of an MS is detected by the BS. Various methods can be used to define and measure signal quality, such as, but not limited to, the RSSI value provided by IEEE 802.11 compliant radio cards.

In principle, the invention consists of a new traffic management method used for the construction of a network composed of different access points (AP), which create an Extended Service Set (ESS), as defined for the IEEE architecture. 802.11, but which act and appear to the terminal as a single Basic Service Set (BSS), also defined in IEEE 802.11, so that, from the point of view of the MS, no handover occurs during its movement from the area of radio coverage of a starting BS to that of a destination BS. During this passage the network behaves, from the point of view of the MS, as a single distributed AP, so that the MS itself, during its movement, even at high speed, believes that it is always connected to the same AP.

With this new technique, a truly “seamless” handover is obtained, meaning that the terminal does not perceive any interruption in the communication both in the transmission and reception phases.

Furthermore, the method we teach, based on a Distributed and Synchronized Control logic, allows network implementations regardless of the topology and transmission means or technologies. Unlike, in part, from the IEEE 802.11 standard (where the ESS is achieved by interconnecting several BSSs through a distribution system (DS) which is normally implemented with a single LAN segment), in our method the BS are connected to form a complete communication infrastructure for backbone transit traffic (hereinafter defined as BCI - Backbone Communication Infrastructure), which can be geographically distributed allowing the creation of resilient networks and of any extension, from a local network (LAN - Local Area Network) to a wide area network (WAN - Wide Area Network).

In one embodiment of the method, a Base Station (BS) is constituted, according to the teachings presented, by the following basic functions:

1. AP emulation,

2. Monitoring of the signal quality of the MS

3. support of the control and coordination logic between the BS for handover

4. traffic management of PDUs to and from the MS and to and from the BCI

5. management of the connection between ESS and external networks.

We call peripheral BS (BBS - Border BS) the one that implements functions 1 to 5 and basic BS or simply BS those that implement functions 1 to 4.

In one implementation the BS is characterized by a “Seamless Handover Controller” (SHC) responsible for managing all the above functions according to the method we teach.

In one embodiment, each BS measures and supervises the signal quality for each individual MS in its cellular coverage area and exchanges this information with all the other BS of the ESS.

The decisions for handover are based, in our method, on monitoring the signal quality of the MS from the point of view of the BS rather than from the point of view of the terminal as prescribed in IEEE 802.11.

The method we teach makes decision-making responsibility distributed in the sense that each BS can decide independently based on the local MS signal quality readings, and those that are asynchronously received by all other BSs.

This way all BS compose a complete view of the signal quality readings for all MS in the network and can make their own decision about sending PDUs to and from each MS.

In the method we teach, it is necessary that the BS converge to a complete common view of the signal quality readings so that the decisions on handover, made by them in a distributed and independent way, tend to be coherent and simultaneous.

In one embodiment, the synchronization function can be obtained by means of a particular BS which always acts as a coordinator, statically configured or dynamically selected within a group of BSs, which periodically transmits its evaluation of the role of each BS for each MS. , or in another embodiment, the synchronization can occur through the use of timestamp information injected into the signal quality messages when all the BS refer to the same time reference through state of the art techniques known as, and not limited to a, NTP (Network Time Protocol).

Regardless of the synchronization technique used, the method we teach works as long as the control data, on which the BS makes decisions, tend to be consistent. But the method employs specific behaviors of the BS to keep the communications of the MS (PDU data traffic) uninterrupted precisely in the periods of time required for the convergence of the control data in the BS.

The method we teach is based on an explicit matrix of BS behaviors (table of forwarding rules shown in Table 1) which guarantee all the exchange flows of the PDUs always operational, from the point of view of a MS. The rules are executed by each BS in a state machine (SM) which defines the roles of the BS for a specific MS. Each BS implements a number of such state machines (one for each MS with which it is exchanging PDUs).

Whenever an event is detected, be it the receipt of a message from the BS designated to the coordinator role, or the expiration of a timer, or the result calculated locally by the SHC on the synchronized quality data, a new role is decided and takes place. a change of state.

Ultimately events cause state changes in the BS and it performs different actions in each state, as will be described in detail below.

In one embodiment, the teachings presented here are also applied to a BS located on the border of the ESS to manage the interconnection to external networks.

This role is achieved through the implementation, on the one hand, of the same SHC functions necessary to manage the PDU traffic to and from the MS in each BS and, on the other hand, through the implementation of a specific forwarding logic from and to to external destinations. The SHC function in fact requires specific behaviors when the PDU traffic concerns external sources or destinations of the communication flow, as further defined in the explicit behavior matrix of the BS. In particular, the method is based on the fact that a border BS knows exactly how and where to forward PDUs when the destination is an MS and that each BS (including border BS), in the traffic management of an MS, knows exactly how and where to forward PDU when the destination address is external to the ESS.

BRIEF DESCRIPTION OF THE DRAWINGS

Figures 1 to 7 describe the PDU flows between one MS and two other MS, while one of them moves from an initial position to a final position within the network. The figures illustrate how the PDU flow transfer functions between a certain number of BSs guarantee a seamless handover, both in the case that all the BSs are based on a common vision of the signal quality reading, as in the case in which they are still converging.

Figure 8 describes the basic network architecture and the minimum number and type of interfaces that are needed in each generic BS and border BS.

Figure 9 describes the state machine which, referring to a specific MS, describes the behavior of a BS.

DETAILED DESCRIPTION

In the following detailed description, numerous specific aspects are set out in order to provide an in-depth understanding of the embodiments of the invention. However, it will be understood by those skilled in the art that embodiments of the invention can be practiced even without these specific details. In other cases, well-known methods, procedures, components and circuits have not been described in detail so as not to obscure the embodiments of the invention.

Throughout this disclosure reference to "a specific embodiment" or "a generic embodiment" means that a particular function, structure or feature described in connection with an embodiment is included in at least one form of the present invention.

Thus, the use of the phrases "in a realization" or "in the realization" in various places of this description are not necessarily all referring to the same embodiment. Furthermore, particular functions, structures or features can be combined in any possible way, in one or more embodiments.

In one embodiment, the described network consists of a series of base stations (BS) that provide access point or "Access Point" (AP) functionality to mobile stations or "Mobile Station" (MS). As described in 801 of Figure 8, a generic BS has a minimum of one physical or logical radio network interface used to exchange PDU with the MS and a minimum of one physical or logical network interface used to interconnect the BS with the backbone network 803 and to allow them to exchange PDUs. The 804 border BS, which implements the "Default Gateway" (DG) function, also has a minimum of an external 805 physical or logical network interface (EI) used for the exchange of PDUs with 802 networks considered external to the ESS.

We call radio interface 806 (RI - Radio Interface), the physical or logical interface that is used in the BS for the exchange of PDUs with the MS and this must be implemented on a radio device.

We call the Backbone Interface 807 (BI), the physical or logical interface used to interconnect BSs and to allow them to swap PDUs. It can be implemented with a radio, optical, electrical connection or any device capable of accomplishing the task.

The method presented here does not depend on how the BS are interconnected to form a backbone network. The method applies to any type of backbone network as long as the physical or logical BI interfaces allow the BS to exchange PDUs transmitted to or from the MS, so that the PDUs can be exchanged over the interfaces (RI-BI ) of the BS both in direct mode and after encapsulation as a payload (or SDU - Service Data Unit) on the specific PDU format implemented on the physical or logical connections of the said BI.

In one embodiment, the BS described here differ from the APs of the IEEE 802.11 standard, of which however they maintain the beaconing and association functions of a standard AP to publish an ESS identifier (ESSID) to the terminals, provide them with basic access. to the network and to allow each of them to associate with a common ESSID, since all the BSs use the same BSSID and the same radio channel on the RI for the exchange of PDUs with the MS. This makes it impossible for the MS to understand which BS they are exchanging PDUs with.

Those skilled in the art may note that, in an ESS standard IEEE 802.11, in which roaming between APs is mandatorily managed in a "break-before-make" mode, said use of the same BSSID and same radio channel in each BS would not work so easily due to unpredictable traffic flows, some of which potentially cause blocking failure, including for example broadcast storms, which could be generated in case of several APs transmitting and, of course, are listening, on the same BSSID and on the same radio channel

While those skilled in the art will certainly recognize that the "emulation of a single BSS" function is easily achievable through the use of the same BSSID, generally identified with the MAC address of the radio device (AP Radio Network Interface or NIC) in each BS, however, they will realize that the method we teach differs from what the IEEE 802.11 standard AP feature does, in the way each individual PDU is handled. In particular, rather than basing PDU forwarding decisions on lists association of STAs (Stations) as standard APs do, we base forwarding and management decisions on explicit forwarding rules that depend on the signal quality metrics of the MS, as further described below.

In our method, all traffic on the RI is blocked by default.

Detailed forwarding actions are defined as numbered rules in Table No. 1.

Although the forwarding actions described by these rules may cause duplicate PDUs to be delivered to a MS, however the handling of such duplicates is outside the scope of our method, and those skilled in the art will recognize that this is usually a function of the upper ISO / OSI layers. As an example, in an IP-type level 3 ISO / OSI implementation, duplicate IP packets are safely discarded at the receiving terminal (duplicate TCP datagrams are deleted at the transport layer, and applications using UDP are called to manage duplicates directly).

On the other hand, the rules of table 1 provide that an MS is always served by at least one BS.

Each forwarding decision is made individually for each PDU.

Thus the traffic management method described in the previous paragraphs and which will be detailed in table 1 below, avoids the problems that standard APs would encounter in this implementation scenario (same BSSID - same radio channel), while the method we teach guarantees that , at any given moment and for each single PDU, there is always at least one BS acting as AP for each single MS.

Which BS is assigned to this task will be decided based on the information collected about the quality of the signal detected at the RI interface, as described below.

The innovative approach to monitoring the quality of the radio signal in order to allow the MS to perform a seamless handover consists in carrying out "network side" monitoring of said signal quality (while IEEE 802.11 is based on a monitoring system of the "terminal side" quality), and in the exchange of such information between each single BS in the network, so that the network as a whole, that is the set of BSs, can get a clear and coherent picture of the signal quality with respect to each single MS and from the point of view of each BS.

In one embodiment, we use the RI to obtain a measure of the signal quality that characterizes the PDUs received by each MS within the coverage area of the cell of a BS; the quality measure can be, in one embodiment, the RSSI (Received Signal Strength Indicator), the noise level, the SNR (signal to noise ratio) or any other measure suitable to characterize the quality of the signal detected by the RI. From now on (and only as an example to describe our method) we will use the term RSSI to refer to a generic measure of signal quality, based on the capabilities of the specific radio device.

Each BS periodically reads on its RI the RSSI of all MS in its cell coverage area and builds a table in which all MS are associated with an RSSI reading. This information is also passed on to all other BSs through BI. Through the exchange of RSSI information, all BSs know exactly how well an MS signal is received by each BS and are able to build an ordered list of "Candidate AP" (ie BS candidates for the role of AP for a specific MS ), on which to make a decision about which BS might be best for serving the MS specification, emulating an IEEE 802.11 AP.

With reference to each specific MS, the following terms are defined:

• Candidate AP - a BS that can detect the presence of an MS in its cell coverage area. A “Candidate AP”, for an MS, can potentially exchange PDUs with it.

• Candidate AP List - A set of “Candidate AP” sorted by signal quality

• Designated AP (DSAP) - the first BS in the “Candidate AP List”.

In one embodiment of the method the BI address of a BS can be the information element that is used to refer to a "Candidate AP".

In each BS, for each MS, it is then possible to decide whether to directly send, through the RI, a PDU to the MS and / or to receive a PDU from the MS or not. In one embodiment, the decision is based on the fact that the BS finds its BI address in the first position of the “Candidate AP List” (in other words it understands that it is the “Designated AP”). If the BS does not find itself as the Designated AP it will still be able to transmit the PDU to the “Designated AP”, but in this case it will use the indirect route (through BI).

Being a distributed control method, it is very robust (it has no Single Point of Failure) and is asynchronous in nature. This implies that the decisions of the BS, due to the continuous movement of the MS between the cellular coverage areas, could be based on signal quality information that is not perfectly updated and coherent with each other, especially since in each BS the local list of " Candidate AP ”is built gradually with asynchronous reception of signal quality updates from all other BSs.

As we will show, the method includes specific PDU management functions which, although based on an ideally synchronized view of quality data, regardless of how the synchronization function is implemented, are expressly designed to ensure the uninterrupted flow of PDUs both when the data quality table is effectively synchronized in all BS, as in a converged view extended to the whole network, and when the signal quality data table, in any BS, is still in the synchronization phase.

Briefly the method we teach is based on the fact that BS persist in forwarding PDUs from an MS even when they are no longer "Designated AP" for it, until it is guaranteed that the signal quality data has been synchronized among all BSs, depending on the particular synchronization method implemented.

Although the implementation of the particular quality data synchronization method is outside the scope of the present invention, it is obtained, by way of non-limiting example, through a particular BS which, always acting as a "coordinator", statically configured o elected dynamically within a more or less extensive group of BSs, and by periodically transmitting to all the other BS his vision of the role of each BS with respect to each MS, he forces their convergence, or that it is obtained, in another example, through "timestamp" information injected into the signal quality messages and forcing the time alignment of all BSs through state-of-the-art techniques such as, but not limited to, NTP (Network Time Protocol), in any case it should be noted that the faster this synchronization occurs or, in other words, the faster the network converges, the shorter the time for which a BS has to keep forwarding PDUs from MS pe r which is no longer the "Designated AP", a state of the BS that we will later call "Dismissed AP" (DAP), and it is obvious that the shorter the residence time of a BS in this state, the lower the number of duplicate PDUs generated to ensure seamless handover.

In one embodiment, as an example relating to the synchronization function, the role of a BS "coordinator" can be introduced with the task of defining, on a T1 time basis, the "Designated AP" for each MS in the network, but it can be replaced, in another embodiment, with any technique suggested by the state of the art for obtaining data synchronization. On the basis of the information gathered about the signal quality, each BS can build a table of "Designated AP" and send it to all the other BS, thus behaving as a BS "coordinator". This coordinator role can be statically configured or can be dynamically elected in a predefined set of BSs. In the case of static configuration of a single “coordinator” it is obviously possible to duplicate it, if this is suggested by reliability considerations. In the case of a BS coordinator elected within a predefined set, a dynamic election algorithm is required. Neither the duplication solution of a single coordinator nor the definition of the election algorithm are within the scope of our method.

By choosing a time base (T1) for the coordinator to periodically send updates that is large (1 second, for example) compared to the transmission delay of control messages within the backbone network, we provide a method of synchronization between all BSs.

Thanks to the method just presented for sharing information on signal quality between BSs, the control logic for handover management can be defined as distributed and synchronized together, in the sense further explained below.

In one embodiment, the method that we present here for a seamless handover, provides that each single BS can assume well-defined states regarding its behavior towards a specific MS at a given instant. In each BS, for each MS, a state machine is implemented as described in 901 of Fig. 9 and is characterized by the definitions of the following states and by the events 902 "New Designated AP detected" and 903 "I am the Designated AP" .

The states are linked to the main flows of the PDUs within each BS. The flows are:

RI-> BI where a PDU enters the BS through the RI and exits through the BI

BI-> RI where a PDU enters the BS through the BI and exits through the RI

RI-> RI where a PDU enters the BS through the RI and exits through the RI

BI-> BI where a PDU enters the BS through the BI and exits through the BI

In the specific case of a Border BS, the RI <-> EI flows have an analogous role to the RI <-> BI flows and the BI <-> EI flows have a similar role to the BI <-> BI flows.

The states are defined as follows, bearing in mind that they always refer to a particular MS:

1. Active AP (AAP)

to. The BS forwards through its BI the PDUs received on the RI by the MS, in other words, it handles the RI-> BI traffic

b. The BS forwards through its RI the PDUs received on the BI and destined for the MS, in other words, it manages the BI-> RI traffic

2. Dismissed AP (DAP)

to. To enter this state, the BS must detect (while in the AAP state) an event that signals that there is a new “Designated AP” for the reference MS.

b. The behavior of a BS in this state (for a particular MS) is described in Table 1 (Table of Forwarding Rules of the BS) and can be summarized as follows:

• For a time T2 the PDUs from the reference MS, entering the RI, are forwarded as if the BS were still in the AAP state.

• the PDUs to the reference MS and received on the RI, are immediately transmitted through the BI to the new "Designated AP" as the forwarding target. In other words, in this case the BS does not have to directly serve the reference MS.

• While the BS is in the "Candidate AP List", even if it is no longer in the first position (ie it is not the "Designated AP"), it forwards through the RI the PDUs coming from the BI and with the target MS. The logic is that if the BS is still in the "Candidate AP List", during the T2 time, it means that it continues to have a connection with the MS and could have a fair chance of reaching it.

• PDUs from the reference MS, received on the RI, destined for another MS for which the BS is in the AAP state, are transmitted directly through the same RI. c. To get out of this state the BS:

• waits for the timer to expire (T2) o

• detects an event indicating that the BS is again the “Designated AP” for the reference MS.

t AP (NAP):

to. The BS does not forward PDUs from the MS through the RI to the BI or RI. Traffic from the MS referral is blocked.

b. The BS transmits through its RI any PDU coming from the backbone and directed to the reference MS, provided that the BS is in the "Candidate AP List" for that MS. The logic is that, for some error conditions, there may be some BS that still thinks this BS is in the AAP state but, if it is still a "Candidate AP", it has a link to the reference MS and a fair probability of reach it anyway.

Therefore, in all these three states the BS forwards on its RI the PDUs coming from the backbone and destined to a MS for which it is still a “Candidate AP” (in other words, the BS still detects the signal of the MS); if, on the other hand, the BS is not in the "Candidate AP List" for that MS (whose signal therefore can no longer be detected by the BS), it will forward the PDUs on the backbone to the real "Designated AP" in an attempt to compensate for the error of the BS of origin and thus increasing the likelihood of delivery of the PDUs to their destination.

Of course in any embodiment of the method it is also possible to block the transmission through the RI of PDUs received from the backbone network, when a BS is in the NAP state for the destination MS, bearing in mind that in this case those PDUs must never be discarded, but transmitted on the backbone towards the BS which is known locally and at that time as “Designated AP” for the destination MS.

It should be carefully noted that this blocking or forwarding of PDUs through BI should only apply to the NAP state of the BS, because otherwise forwarding loops can occur when a BS in DAP state for a certain MS receives a PDU for this MS and forwards it over the backbone network to the BS which is known locally as “Designated AP” at the time of the decision, but which has not yet received the update which makes it an AAP for that MS.

The BS forwarding rules used in our method are detailed in Table 1.

We define RS (remote station), any PDU destination external to the ESS network, reachable via a border BS connected to the network outside the ESS and acting as a gateway in any layer of the specifically implemented ISO / OSI model, such as, and not limited to, an IP Gateway (Level 3) or an Ethernet bridge (Level 2), henceforth conceptually called Default Gateway (DG).

From the point of view of a BS we distinguish between different states in which an MS may be, while noting that the MS is completely unaware of these states.

The states are:

• MMS (Managed Mobile Station): the BS is able to detect the signal of the MS and is in AAP state for that MS

• DMS (Departing Mobile Station): the BS is able to detect the signal of an MS and is in DAP state for that MS

• IMS (In-range Mobile Staion): the BS can detect the signal of an MS, but it is in NAP state for that MS

• OMS (Out-of-range Mobile Station): the BS is not able to detect the signal of an MS but it is announced by at least one BS on the backbone network.

All destinations that are not within the ESS are considered remote stations or “Remote Stations” (RS), therefore potentially reachable through the DG.

In the BS forwarding rules table (Table 1) the definitions of the column headings are as follows:

• Input interface: identification element for the interface from which the PDU enters the BS

• Origin of the PDU: identification element for the origin of the PDU • Destination of the PDU: identification element for the final recipient of the PDU

• Decision of the BS: describes the action that the BS performs

• Output interface: identification element for the interface from which the PDUs leave the BS

• Forwarding target: identification element for the destination of the forwarding action of the BS.

The "Duplicate Route" column is not needed to define a rule. It is given only to describe whether the rule refers to the main path (only one copy of the PDU is typically delivered to the MS) or to a path that may cause duplicate PDUs at the destination (the MS may receive duplicate copies of the PDU). In rules 11, 12, 13, 14 and 15 the MS will not receive any PDU (it is assumed that there is at least one other BS managing the PDUs arriving from that MS at that time).

All the rules that the method we teach must implement to manage the communication flows between MS and between an MS and an RS are described with reference to these definitions and are listed as numbered rules in table 1. In this table an MMS which is also the destination of a PDU is indicated with “dMMS” (destination MMS).

Table 1 BS forwarding rules

Rules 5bis, 10bis, 21bis, 27bis, 33bis and 39bis apply only to a border BS and replace the corresponding rules 5, 10 21, 27, 33 and 39 referring to a generic BS. In these rules the forwarding target does not apply, as the forwarding decision is based only on the outbound interface (which is EI).

All other rules, starting from number 1 and up to number 45, apply to both generic and border BSs.

Rules 46 to 51, of course, only apply to the border BS.

Rules 18/19, 24/25, 30/31, 36/37, 42/43 and 48/49 are mutually exclusive, meaning that neither is mandatory but at least one must be implemented. The choice is left to the specific network configuration.

Those skilled in the art will readily recognize that rules 1 to 15 implement the previously described function of "emulating a single BSS", avoiding the related potential blocking failure events of the network.

In order for the method we teach to work properly, a minimum of PDU traffic is required from each single MS; at a minimum we need a PDU from each MS in each time slot we call T3.

The description of traffic generation techniques is outside the scope of the method presented here. In one embodiment, any state of the art technique can be used to achieve the required result. As an example, in an IP-type ISO / OSI Level 3 implementation, periodic ARP requests could be used to explicitly solicit ARP responses from any MS that supports the IP protocol in the event of no spontaneous PDU traffic inside. time interval T3.

Depending on the type of network installed, the timings T1, T2 and T3 can assume different values. Under the condition T2> T1> T3 the values can be optimized to obtain the best performance in handover.

DESCRIPTION OF HANDOVER DRAWINGS

For a detailed description of an embodiment of the invention, reference is now made to the drawings of figures 1 to 7. In these figures the network backbone is represented with a bus for convenience, but any network topology could apply.

In 108, 208, 308, 408, 508, 608 and 708 each row represents a "Candidate AP List" for a particular MS, and we assume for simplicity that it consists of only one element, which obviously identifies the "Designated AP". This is sufficient for the "Standard" operating mode described up to now for the SHC. The general case of a "Candidate AP List" that includes more than one element will be considered below, to illustrate the most advanced operating methods for the SHC, respectively called "Redundant", "Fully Redundant" and "Optimized Fully Redundant" ".

Each figure represents a moment in time during the handover process. In particular:

• a T0: In figure 1, being 103 the backbone network, 104a, b, c, d the coverage areas of the cells identified with the BSS, 105 the ESS, 106 the time axis (on which the spatial intervals do not are proportional to the duration of the time intervals, which cannot be known in advance), 107a, b, c, d the information sent by each BS about the quality of each signal detected by the MS, the MS 101b is in the coverage area of the cell 104b of BS 102b. Only 102b is capable of detecting the signal and reports the signal quality reading (RSSI) to all other BSs (102a, 102c and 102d). With these ratios all the BS can build a table like 108. The PDUs exchanged between 101a and 101b follow the path 109a, 109b and 109c.

The PDUs exchanged between 101b and 101c follow the path 110a, 110b and 110c • at T1: in figure 2, 201b enters the coverage area of cell 204c of BS 202c. BS 202c detects the signal from MS 201b and reports the signal quality reading to all other BSs (202a, 202b and 202d). As indicated in Table 207c, this reading in 202c is lower than the signal quality reading for MS 201b to BS 202b (see Table 207b). This is reflected in a table like 208 that each BS can build, as it also collects all the other readings from the various BSs. In fact, Table 208 states that BS 202b deals with MS 201b (202b is "Designated AP" for 201b).

• at T2: in figure 3, the signal of MS 301b is detected with almost the same RSSI by the BS 302b and 302c and this is reflected in tables 307b and 307c. In general, to avoid fluctuations in the construction of boards like the 308, a threshold could be used in the decision making process. All BS 302a, 302b, 302c and 302d are assumed to be aligned (i.e. synchronized) to the view expressed in Table 308. The streams of PDUs 309a, 309b and 309c and 310a, 310b and 310c express the fact that each BS treats 302b as the " Designated AP "for MS 301b and that 302b is in the AAP state for 301b.

• a T3: In Figure 4, the MS 401b signal is better detected by BS 402c instead of 402b and its RSSI readings in 402b and 402c are depicted in Tables 407b and 407c respectively. The handover of 401b from 402b to 402c is reflected in table 408. If not synchronized, this handover decision (ie the construction of 408) occurs asynchronously in each BS. As soon as 402b realizes that it must leave the AAP state for 401b it enters the DAP (Dismissed AP) state for 401b and the T2 timer is started, as described in Figure 9. During this period, 402b remains in the DAP state if it detects no more. events. In figure 4 it is assumed that at T3BS 402d has converged on the common view expressed by table 408, but that 402a is still in the process of convergence towards 408, so as to still believe that 402b is the "Designated AP" for 401b (in essence it maintains the view expressed in the old table 308). This situation produces the PDU streams as represented by 409a, 409b, and 409c. It is the result of the fact that the BS 402b, 402c and 402d are synchronized on the opinion expressed by the 408, while the 402a has not yet completed the convergence phase. Flows 410d, 410e and 410f are the result of the convergence of 402c and 402d on table 408. Flows 409f, 409e and 409d are due to 402c behaving like AAP for 401b, but the flow is unidirectional due to the fact that 402a is not yet synchronized on 408. The flows 410a, 410b and 410c are due to the fact that 402b is in the DAP state for 401b. It should be noted that the behavior of BS 402b would have been exactly the same even if at T3 it was not yet converted to 408, in other words if it still behaved as AAP for 401b.

• A T4: in figure 5 it is assumed that the BS 502a is finally synchronized with 508, as represented with the traffic flows 509f, 509e and 509d for the traffic 501a-501b. BS 502b is still waiting for the T2 timer to expire, which means it is still in DAP state for 501b, and then forwards PDUs from 501b to the network as illustrated in 509c, 509b and 509e, and in 510a, 510b and 510c, thereby generating duplicate one-way PDUs.

• to T5: in figure 6 finally it is assumed that all BS (601a, 601b, 601c and 601d) are synchronized on the same vision for their respective roles (see 608). In this case 602a, as all other BS do, forwards the PDUs for 601b to 602c (the now commonly recognized "Designated AP" for 601b). This completes the considered handover process.

• at T6: in figure 7, 701b is now detected only by 702c (for which BS is an MMS) and is an OMS for 702b.

The sequence of events described in Figures 1 to 7 is just an example. By applying the rules of table 1, all possible cases can be analyzed and it can be shown that there can be duplication of PDUs in the communication flows between MS, but it never happens that there is no BS serving any MS in a given moment.

The use of a coordination function, however implemented and as already described, helps the asynchronous convergence of the single local visions of reachability of the MS in each BS, forcing at defined intervals T1 <T2 the adoption of a common and coherent framework of the roles of BS for each single MS, extended to the whole network.

The case of traffic flows between an MS and an RS is easily analyzed by replacing one of the BS in Figures 1 to 7 with a boundary BS and applying the rules of Table 1 again.

In the example described with figures 1 to 7, the simple case of a "Candidate AP" table consisting of a single AP column is considered, but this case represents only the minimum information required and can be extended to the general case of a table with n number of "Candidate AP" columns sorted from left to right in ascending order of priority.

With a table extended in this way the following modes or modes of operation can be defined and used:

1. Standard: in each BS, the PDU streams delivered to the backbone network are sent to the BS alone in the AAP state (the single "Designated AP") for a particular MS, so that there can be no duplicate PDU paths from the network dorsal to the MS. PDU flows from MS to backbone are served by BS which are in both AAP and DAP state until T2 expires. This can cause duplicate PDU flows in the MS-to-backbone direction, but the MS are always served during transmission.

2. Redundant: PDU flows for a specific MS are sent, as well as to the Designated AP, also to the BS which is in the "Dismissed AP" state (the one that had been in the "Designated AP" column in previous T1 cycles, and which could now be in the second position of the "Candidate AP List" for a specific MS), in addition to what happens in the Standard mode, in which the PDU flows from a specific MS are also managed by the BS which is in the status of " Dismissed AP ”, just like the Designated AP does. During the convergence time T1 <T2, duplicate paths and consequently duplicate PDUs are thus created, also from the backbone network to the reference MS.

3. Fully Redundant: in the backbone to MS direction a number n of routes and consequently (n-1) duplicate PDUs are created to reach a specific MS, n being the level of route redundancy required for a particular method implementation . In the MS to backbone direction, all BSs that are within the first n "Candidate APs" for a specific MS will forward the PDUs from the MS to the backbone.

Those skilled in the art will recognize that the timer T2 which defines the permanence of a BS in the DAP state before passing to the NAP state, is necessary, in the state machine 901 of Figure 9, only for the Standard and Redundant operating modes with a single column of "Candidate AP" in tables 108-708 (the so-called "Designated AP") as, in Fully Redundant mode, the only event that must cause a transition to the NAP state in the BS (directly from the AAP state) is that it finds itself in (n + 1) <-th> position in the relative "Candidate AP List" (ie the row corresponding to the reference MS in tables 108 -708), where n is the path redundancy level required.

The Fully Redundant operating mode can be used, at the cost of a higher level of PDU traffic in the backbone network, in those situations where there is a need to completely avoid PDU loss. In this operating mode, all possible paths to reach an MS are attempted (up to n) and all access points that have the possibility to reach the MS will try to serve it.

By way of example, in an implementation of the method in Fully Redundant mode we can assume n = 2 and, for each PDU transmitted from one MS to another MS, when both belong to the same ESS, there will be 4 copies of each PDU ( in other words three duplicates). The first two copies are due to the first 2 BS of the “Candidate AP List” for the MS of origin that “serve” the MS, in fact each of the two BS forwards a copy of the PDU. The second two copies are created as each of the 2 BS that receives the PDU will forward it to the first 2 BS of the "Candidate AP List" for the destination MS, thus maximizing the probability of maintaining the communication always functional, obviously in the face of a certain cost in terms of transmission bandwidth on the network due to the overhead of duplicated traffic.

It should be noted that, even if the Standard operating mode is already sufficient, thanks to the properties of the forwarding rules described in table 1, to ensure, at least in theory, that the flow of PDUs is always without interruptions during the handover phase, this may not be the case in harsh environments characterized by strong disturbances and a consequent high packet loss due to transmission errors. These phenomena, if present, are particularly relevant near the boundary of a coverage area, where the MS signal is normally the weakest for both BSs involved in the handover process. Obviously, the Fully Redundant mode can be of great use even in the case of a high rate of loss of PDUs within the backbone network that interconnects the BSs.

This mode of operation can be further improved to offer multiple paths to PDUs only during the handover phase between BSs, and not during periods when the MS is static or far from the boundary of the cell coverage area for a BS. In an implementation using a FIFO (First In First Out) storage mode of the historical data collected to build each "Candidate AP List", monitoring them as they vary over time, it is possible to better understand the behavior of a specific MS when it moves between the cells, so that it becomes easy to understand when an MS is at the limit of the cell coverage area for its "Designate AP", which means that a handover between BS is about to begin with the consequent occurrence of an event that causes a BS from Standard to Fully Redundant mode when a certain threshold is passed, and then returns to the first when the event is over.

In fact, if the signal value is used to describe the difference of the signals perceived when a MS reaches the overlapping zone of the cell covers of two BS, as in 305b and 305c where the difference in the signals perceived by the BS 302b and 302c, for 301b has a value of 1, after setting a threshold value for the difference, it is easy to recognize a phase in which the MS 301b is potentially about to generate a handover shortly. Only then, the forwarding behavior of the "Candidate AP" could be modified in order to generate more routes to and from the MS, thus maximizing the chances of delivery of the PDUs before the handover process actually begins.

If a specific MS is found to be far from the handover condition, its "Candidate AP List" (a row in tables 108-708, when they have multiple AP columns) can be used to extract only the first element, as in the way of Standard operation, in which case, the behavior of the BS reverts to Standard mode.

In one embodiment of the method we call this "Optimized Fully Redundant Mode" and it is simply defined as a Standard operating mode, if the difference in perceived signals for a MS is above a certain threshold, but with the possibility of passing to the "Fully Redundant" mode for a particular MS, when said threshold is reached.

This dynamic behavior can provide a very low probability of PDU loss, even in harsh environments, along with efficient utilization of network resources.

QUOTES

1 - [Bejerano et al. 2002] BEJERANO, Y., CIDON, I., AND NAOR, J. S. 2002. Efficient handoff rerouting algorithms: A competitive on-line algorithmic approach. IEEE / ACM Trans. Netw. 10, 6, 749–760.

2 - [Caceres and Padmanabhan 1998] CACERES, R. AND PADMANABHAN, V. N. 1998. Fast and scalable wireless handoffs in support of mobile internet audio. ACM J. Mobile Netw. Appl. 3, 4, 351–363.

3 - [Campbell et al. [2002] CAMPBELL, A., GOMEZ, J., KIM, S., WAN, C.-Y., TURANYI, Z., AND VALKO, A. 2002. Comparison of ip micromobility protocols. IEEE Trans. Wirel. Comm. 9, 1, 72–82.

4 - [Chiasserini 2002] CHIASSERINI, C.-F. 2002. Handovers in wireless ATM networks: In-Band signaling protocols and performance analysis. IEEE Trans. Wirel. Comm. 1, 1.

5 - [Das et al. 2002] DAS, S., MCAULEY, A., DUTTA, A., MISRA, A., CHAKRABORTY, K., AND DAS, S. 2002. IDMP: An intradomain mobility management protocol for next-generation wireless networks. IEEE Trans. Wirel. Comm. 9, 3.

6 - [Ganguly et al. 2006] GANGULY, S., NAVDA, V., KIM, K., KASHYAP, A., NICULESCU, D., IZMAILOV, R., HONG, S., AND DAS, S. 2006. Performance optimizations for deploying voip services in mesh networks. IEEE J. Select. Areas Comm. 24, 11, 2147-2158.

7 - [Grilo et al. 2001] GRILO, A., ESTRELA, P., AND NUNES, M. 2001. Terminal independent mobility for ip (timip). IEEE Comm. Mag. 39, 12, 34–41.

8 - [Helmy et al. 2004] HELMY, A. A.-G., JASEEMUDDIN, M., AND BHASKARA, G. 2004. Multicast-Based mobility: A novel architecture for efficient micromobility. IEEE J. Select. Areas Comm.

9 - [Hsieh et al. 2003] HSIEH, R., ZHOU, Z. G., AND SENEVIRATNE, A.

2003. S-MIP: A seamless handoff architecture for mobile IP. In Annual Joint Conference of the IEEE Computer and Communications Societies (INFOCOM).

10 - [Mishra et al. 2003] MISHRA, A., SHIN, M., AND ARBAUGH, W. 2003. An empirical analysis of the IEEE 802.11 MAC layer handoff process. SIGCOMM Comput. Comm. Rev. 33, 2, 93-102.

11 - [Navda et al. 2005] NAVDA, V., KASHYAP, A., AND DAS, S. 2005. Design and evaluation of imesh: An infrastructure-mode wireless mesh network. In 6th IEEE WoWMoM Symposium.

12 - [Perkins 1996] PERKINS, C. 1996. IP mobility support. RFC2002.

13 - [Ramachandran et al. 2005] RAMACHANDRAN, K. N., BUDDHIKOT, M. M., CHANDRANMENON, G., MILLER, S., BELDING-ROYER, E. M., AND ALMEROTH, K. C. 2005. On the design and implementation of infrastructure mesh networks. In Proceedings of the IEEE Workshop on Wireless Mesh Networks (WiMesh).

14 - [Ramani and Savage 2005] RAMANI, I. AND SAVAGE, S. 2005. Syncscan: Practical fast handoff for 802.11 infrastructure networks. In Proceedings of the Annual Joint Conference of the IEEE Computer and Communications Societies (INFOCOM).

15 - [Ramjee et al. 1999] RAMJEE, R., LA PORTA, T., THUEL, S., VARADHAN, K., AND WANG, S. 1999. Hawaii: A domain-based approach for supporting mobility in wide-area wireless networks. In Proceedings of the 7th International Conference on Network Protocols (ICNP'99). 283–292.

16 - [Seshan et al. 1996] SESHAN, S., BALAKRISHNAN, H., AND KATZ, R.

1996. Handoffs in cellular wireless networks: The Daedalus implementation and experience. Kluwer J. Wirel. Personal Comm

17 - [Sharma et al. 2004] SHARMA, S., ZHU, N., AND CKER CHIUEH, T.

2004. Low-Latency mobile ip handoff for infrastructure-mode wireless lans. IEEE J. Select. Areas Comm. 22, 4, 643–652.

18 - [SMESH. 2010] The SMesh wireless mesh network. http://www.smesh.org.

19 - [Soliman and Bellier 2004] SOLIMAN, H., CASTELLUCCIA, C. K. M., AND BELLIER, L. 2004. Hierarchical mobile ipv6 mobility management (hmipv6). IETF RFC 4140.

20 - [Valko 1999] VALKO, A. G. 1999. Cellular ip: A new approach to internet host mobility. SIGCOMM Comput. Comm. Rev. 29, 1, 50–65.

21 - [Vatn 2003] VATN, J.-O. 2003. An experimental study of IEEE 802.11b handover performance and its effect on voice traffic. Tech. rep. TRITA-IMIT-TSLAB R 03:01, Department of Microelectronics and

Information Technology, KTH, Royal Institute of Technology, Stockholm, Sweden.

22 - [Yokota et al. 2002] YOKOTA, H., IDOUE, A., AND HASEGAWA, T.

2002. Link layer assisted mobile ip fast handoff method over wireless lan

<5> networks. In Proceedings of ACM / IEEE International Conference on Mobile Computing and Networking (MOBICOM). ACM Press, New York, 131–139.

Claims (9)

CLAIMS Claims of the invention having as title: "Method and Apparatus for Seamless Handover" network side ", invisible" terminal side ", for terminals compliant with the IEEE 802.11 standard, based on a distributed and synchronized handover control logic, regardless of the type of network backbone" CLAIMS Indeed, the present invention is not limited to the foregoing discussion and accompanying drawings. Rather, the present invention is limited only by the following claims and their legal equivalents and it must be understood by those skilled in the art that the communication method proposed together with the apparatus that allows an unmodified IEEE 802.11 WiFi standard terminal to perform a seamless handover, as indicated in this invention, can be modified considerably without departing from the spirit and scope of the invention as defined by the following claims. The following is claimed: 1. A method, in a wireless network comprising, regardless of the type of backbone interconnection, several Base Stations (BS), which play the role of Access Point (AP) to create an Extended Service Set (ESS), as defined in IEEE 802.11, where these APs exchange Protocol Data Units (PDUs) with Terminals compliant with the IEEE 802.11 standard, to support, for these Terminals, the Seamless Handover "network side", invisible "terminal side", when one of them moves from an Access Point to another, such method comprising: emulation function of a single IEEE 802.11 Access Point for the entire network, through the use of a single BSSID on each BS of the wireless network, used to implement an ESS that acts and appears to the terminal as a single Basic Service Set (BSS) so that, from the perspective of the terminal, no handover occurs while it moves from an initial AP to a final AP, as said terminal believes it is connected to the same AP all the time. making decisions on the handover of the terminal independently in each BS, basing these decisions on the reading of the signal quality of the terminal carried out by each BS at the level of its radio interface (RI) and shared, among all the other BSs, through the connection to the network backbone, to obtain a clear and consistent picture of the signal quality with respect to each individual terminal and from the point of view of each BS in the network. PDU management functions which, although based on a synchronized view of the signal quality data, are expressly designed to ensure the uninterrupted flow of the PDUs both when the signal quality data table is synchronized, as in a convergent vision extended to the whole network, both when the signal quality data table, in any BS, is still being synchronized the list of explicit rules on which the management functions of the PDUs base their forwarding decisions, in both "terminalerete" and "network-terminal" directions, always on the basis of the individual PDU the list of explicit rules on which the PDU management functions base their forwarding decisions, always on the basis of the single PDU, in the particular case of a "border" BS that interconnects the ESS and any external network the “Redundant”, “Fully Redundant” and “Optimized Fully Redundant” operating modes for increasing reliability in the delivery of PDUs, in addition to the basic behavior of the BS. 2. The method, according to claim 1, where the emulation of a single BSS at the level of the entire ESS, includes providing beaconing and association functions of a standard AP to publish an ESS identifier (ESSID) to the terminals, provide them with basic access to the network and allow each of them to associate with a common ESSID forcing the radio functions, such as beaconing, associations, and in general transmission and reception, to be performed in the same identical wireless radio channel for the entire network the blocking of all the forwarding functions of the PDUs typical of a standard AP, in order to avoid unpredictable traffic flows, some of which potentially cause a blocking failure, including for example broadcast storms, which could be generated in the case of several APs that transmit and, of course, are listening, on the same BSSID and on the same radio channel. The method according to claim 1, wherein the use of the signal quality reading to build a quality data table, comprises the use of the radio interface of each BS to read the signal quality of each terminal in the coverage area the creation, in each BS, of a table of signal quality measures for each terminal in the coverage area the sharing of these tables among all the BSs so that each of them is able to create a synchronized table containing all the terminals in the coverage area of the entire ESS and the related signal quality measures for each terminal at each BS the creation of a list of BS, ordered according to the strongest signal measured for each terminal, candidates to manage the traffic of PDUs to and from said terminal the use of this list to make decisions regarding the handover of a terminal, in terms of which BS is from time to time responsible for managing the traffic of PDUs to and from said terminal. The method according to claim 1, wherein the PDU management functions include locally in each BS, as soon as the signal quality data table is available in relation to a terminal, the decision as to whether or not a BS should be in charge of forwarding the PDUs from that terminal, and which BS is to be choose for forwarding the PDUs destined for that terminal in any BS that realizes that it is no longer the best candidate to handle PDU traffic for a particular terminal, basing this assumption on the last available signal quality data table, continuing to transmit PDUs from said terminal to the network and to deliver PDUs eventually received from the network and destined to said terminal, for a configured time longer than the time typically foreseen for the propagation throughout the network, in order to create a synchronized view for all the BSs, of the quality data of the signal, thus obtaining, with this technique, an explicit method to guarantee an uninterrupted flow of PDUs and, therefore, from a terminal point of view, a seamless handover precisely during the entire period of time in which the quality data table of the signal is not yet fully synchronized, regardless of how and with which technique such synchronization is achieved. 5. The method, according to claim 1, where the list of explicit rules on which the management functions of the PDUs base their forwarding decisions, includes explicit rules for deciding, on the basis of the single PDU, whether or not to forward a PDU received on the radio interface, basing this decision on the sending terminal and on the previous evaluation of the signal quality data relative to this terminal, as in the previous claim 4 explicit rules for deciding, on the basis of the single PDU, if already selected for transmission, to where to forward a PDU received on the radio interface, basing the decision on the destination terminal and on the previous evaluation of the quality signal data relative to this terminal , as in the previous claim 4 explicit rules for deciding, on the basis of the single PDU, towards where to forward a PDU received on the backbone network interface, basing the decision on the destination terminal and on the previous evaluation of the signal quality data relative to that terminal, as in the previous claim 4. 6. The method of claim 1, in which the list of explicit rules on which the PDU management functions are based, in the special case of a border BS that interconnects the ESS with an external network, includes explicit rules for deciding, on the basis of the single PDU, whether or not to forward a PDU received on the radio interface, basing this decision on the sending terminal and on the previous evaluation of the signal quality data relative to this terminal, as in the previous claim 4 explicit rules to decide, on the basis of the single PDU, whether or not to forward a PDU received on the radio interface and already selected for transmission, basing the decision on the destination of the PDU, be it the external network or the internal network, and, in the latter case, basing the decision on the destination terminal and on the previous evaluation of the data of the quality signal relative to that terminal, as in the previous claim 4 explicit rules to decide, on the basis of the single PDU, whether or not to forward a PDU received on the backbone network interface, basing the decision on the destination of the PDU, be it the external network or the internal network, and in the latter case basing the decision on the destination terminal and on the previous evaluation of the quality signal data relative to that terminal, as in the preceding claim 4 explicit rules for deciding, on the basis of the single PDU, whether or not to forward a PDU received on the external network interface, and destined to a terminal in the internal network, basing the decision on the destination terminal and on the previous evaluation of the signal data quality relative to this terminal, as in the previous claim 4. 7. The method of claim 1, where the operating modes "Redundant", "Fully Redundant" and "Optimized Fully Redundant" for increasing reliability in the delivery of PDUs, in addition to the basic behavior of the BS, include in the "Redundant" mode, when a BS is no longer the best candidate for PDU traffic management for a specific MS, according to the most recent available signal quality data table, duplicate the PDUs destined for that MS, delivering the PDUs originals to said MS and those duplicated to the new best BS candidate for PDU traffic management for said MS, maintaining this behavior for a longer time than the expected time of propagation of quality data through the network in the "Fully redundant" mode regardless of the expected time of propagation of quality data through the network, the PDUs are always duplicated both in the direction from network to terminal and vice versa and each single copy of the PDU is sent to the first two candidates for traffic management for a specific terminal, so as to maximize the probability that at least one copy reaches the terminal and in the terminal-to-network direction, allowing both said candidates to forward the PDUs sent by the terminal at the same time, where the term "duplicate" can be replaced to be "multiplied by a given whole number" again in order to maximize the probability that at least one copy reaches the terminal in the "Optimized Fully Redundant" mode, a behavior like that of the "Fully Redundant" mode but, referring to a particular terminal, only when said terminal approaches the boundary of the cell or is about to meet the handover conditions and a behavior like that of the Standard operating mode when these conditions are no longer verified. 8. A base station or "Base Station" (BS), in a wireless network comprising several BS, which play the role of Access point (AP) to realize an Extended Service Set (ESS), as defined in IEEE 802.11, where such AP exchange Protocol Data Units (PDUs) with Terminals compliant with the IEEE 802.11 standard, to support, for these Terminals, the Seamless Handover "network side", invisible "terminal side", when one of them moves from one Access Point to another , the base station comprising a minimum of one physical or logical wireless network interface used to exchange PDU terminals with mobile terminals a minimum of one physical or logical network interface used to interconnect the BSs and to allow them to exchange PDUs. a Seamless Handover Controller (SHC) configured to: provide IEEE 802.11 Access Point emulation functionality throughout the network, through the use of a unique BSSID on each BS, in order to implement an ESS that appears to the terminal as a unique Basic Service Set (BSS), and such that, from the point of view of the terminal, no handover occurs while it passes from an initial AP to an arrival AP, as said terminal believes that it is connected to the same and unique AP all the time make handover decisions in each BS independently, basing them on the terminal signal quality measurements made by each BS at its radio interface level and sharing the measurements among all other BS to get a clear and consistent picture of the quality of the signal related to each single terminal and from the point of view of each BS in the whole network perform PDU management functions, based on a list of explicit rules relating to both the terminal-network direction and the network-terminal direction, always for each individual PDU. 9. A Border Base Station or "Border BS", in a wireless network comprising several Base Stations (BS), which play the role of Access Point (AP) to create an Extended Service Set (ESS), as defined in IEEE 802.11 , where these APs exchange Protocol Data Units (PDUs) with Terminals compliant with the IEEE 802.11 standard, to support, for these Terminals, the Seamless Handover "network side", invisible "terminal side", when one of them moves from an Access Point to another, and which acts as an interconnection element between the ESS and an external network, the border base station comprising: a minimum of one physical or logical wireless network interface used to exchange PDUs with terminals a minimum of one physical or logical network interface used to interconnect the BSs and to allow them to exchange PDUs a minimum of one physical or logical network interface used to exchange PDUs with networks considered external to the ESS a Seamless Handover Controller (SHC) configured to: provide IEEE 802.11 Access Point emulation functionality throughout the network, through the use of a unique BSSID on each BS, in order to implement an ESS that appears to the terminal as a unique Basic Service Set (BSS), and such that, from the point of view of the terminal, no handover occurs while it passes from an initial AP to an arrival AP, as said terminal believes that it is connected to the same and unique AP all the time. make handover decisions in each BS independently, basing them on the terminal signal quality measurements made by each BS at its radio interface level and sharing the measurements among all other BS to get a clear and consistent picture of the quality of the signal related to each single terminal and from the point of view of each BS in the whole network. perform PDU management functions, based on a list of explicit rules relating to both the terminal-network direction and the network-terminal direction, always for each individual PDU. perform PDU management functions, based on a list of explicit rules, always based on individual PDUs, in the particular case of a border BS that interconnects the ESS with any external network.