Introductory Scenarios - Hovering Information Model

Hovering Information Model

3.1 Introductory Scenarios

In order to present and describe the hovering information model, we will make use of two use case scenarios that we will develop progressively throughout the entire chapter in order to illustrate the different concepts making part of the model. The following subsections introduce two general social scenarios: LostKeys and FleeMarket.

3.1.1 Lost Keys

At a New Year’s music and dance festival, Bob who is attending the festival with his friends is wandering around, looking for an appealing scene. As it is usual in this kind of festivals, the place is crowded and moving through the crowd might be difficult at times. However, even though it is a struggle finding/fighting his way through all the people, Bob is enjoying the festival until he notices that he lost his keys! As he is sure he had them when arriving at the festival, he is worried that he lost them while walking around in the crowd. Unfortunately, there is not even

“lost & found” objects office at the festival! How should Bob find them?

A natural and common reaction is trying to remember one’s own path or route and walking it back step by step. However, due to the amount of people around, it is certainly difficult to walk along a specific way while looking carefully at the ground in the hope of finding the keys.

In the meantime, Tom who is also enjoying the festival and does not know Bob, finds the keys, laying down in the floor. He wonders whether he should try to find a “lost & found” office but he is not sure whether there is one at the festival. And even if there would be one, the big crowd of people would make reaching the office a difficult task. Also, together with his friends

Tom was just looking for a stage where a nice concert was supposed to start in a few minutes.

What should he do?

Fortunately, both Bob and Tom have the LostKeys mobile application installed in their mo-bile phones. Tom, after having found the keys, puts up a short message at the place where he found the keys. The message says:“I found a set of keys here. If you want them back contact me at....”. Then he leaves the place where he found the keys to continue enjoying the festival. Bob decides to walk back the way he came from, searching for his keys. But to improve his chances of finding them, specially in such a crowded place, he also uses the LostKeys mobile application that enables him to see all the messages that people attached to specific locations when having found keys.

It might also happen that Bob looses his keys around midnight. As it is New Years eve and the place is extremely crowded, it is highly probable that all mobile networks will be overloaded.

Fortunately, the LostKeys mobile application does not rely on the cellular infrastructure but on the ad hoc networking capacities of mobiles phones, bypassing cellular antennas.

Still in search for his keys, Bob passes nearby the place where he lost his keys and the messages Tom left shows up on his mobile phone. He decides to contact Tom by replying to the message. Depending on whether or not Tom left his mobile phone number as the contact information, Bob can either send him a text message through the mobile operator or reply to the message by using the LostKeys application. In the latter case, which would be preferable on New Years eve as the network might still be overloaded, Bob puts up a response message to the entire festival area as he does not know where Tom is currently located. However, unlike the message that Tom posted, Bob’s reply message has Tom as the only target person instead of everyone that posted a “lost & found” keys message. Tom eventually receives the message and replies by again putting up a reply message to the whole festival area and destined to Bob.

They agree on meeting at the “Fish & Chips” kiosk next to the big stage where Tom is enjoying a great concert.

Finally, Bob joins Tom at the kiosk and gets his keys back! Tom may now decide to remove the initial message he posted or just leave it until it will disappear at the end of the festival.

Figure3.1illustrates the described scenario.

3.1.2 Flee Market

Alice is going shopping at a huge and crowded flee market (like a bazaar). Among other kitchen equipment she is looking for a toaster and a mixer that she needs for her house. However, as the number of offers is huge, it is not easy to find the place with the best quality-price rate, also offering a polite service. She decides to use the FleeMarket mobile application she installed in her mobile phone to facilitate her selection process. She configured the application to make her phone vibrate whenever she is nearby an interesting seller. But how does the application know that such seller is better than others?

Other people walking around also make use of the FleeMarket to ease their shopping. One of them is Fred who is also looking for kitchen equipment. He just bought a mixer and he is very happy with the kindness of the seller and price of the product. So he decided to just put up a review message at the place of the seller’s kiosk saying: “This is a nice place where to buy kitchen equipment. Good price. Good quality. Good service” and he leaves. As Fred, many other people do the same, producing thereby a lot of review messages. It is up to the FleeMarket application to collect all the reviews for a place and combine them to produce a global review.

(a) Bob notices he lost his keys

(b) Tom find the keys (c) Tom puts up a mes-sage at the place where he found the keys

(d) Bob walks back looking carefully for his keys and finds the message that Tom left

(e) Bob contacts Tom to get his keys back by putting up a message destined to Tom at the whole festival area

(f) Tom replies to Bob by putting up a mes-sage destined to Bob at the whole festival area.

He proposes him to meet at the “Fish & Chips”

kiosk to give him his keys back.

Figure 3.1: LostKeys scenario

It is based on the global review that a mobile phone notifies or not his user accordingly to his interests.

When Alice is passing by the place where Fred bought his mixer, her mobile phone vibrates as that place has a good global review. Alice decides to buy the toaster and mixer that she was looking for at that seller’s place. And she also leaves a review message when leaving so that other people may benefit from her experience.

3.2 Components

We distinguish six components making part of the hovering information paradigm. Following a bottom-up approach, we first identify themobile nodes, which can be any mobile device on which hovering information relies to exist. Then we have thehovering informationitself which is anchored to a geographical area. The third component are thereplicaswhich are the instances of hovering information residing in the mobile nodes. Thehovering information serviceenables mobile applications to make use of hovering information by making abstraction of the mobile nodes and replicas and offering primitives to storage and retrieve hovering information into/from geographical regions. Finally,mobile usersrepresent people or systems making use of mobile applications in order to achieve a precise task. The following subsections define each of these components in detail.

3.2.1 Mobile Nodes

A mobile nodecan be any mobile device provided it has computing, storage and networking resources which are indispensable and exploited by hovering information for its existence. There exists a great variety of mobile nodes due to the large number of different types of mobile devices. For instance, the market offers a selection of mobile phones and laptops, each of them having specific characteristics in terms of storage capacity or networking interfaces among others. In the LostKeys and FleeMarket scenarios, we can distinguish as mobile nodes all the different types of mobiles phones carried by people at the festival or the flee market, as well as the laptops used by the organisers of the festival or some sellers.

Formally, we can define a mobile nodenas an n-tuple:

n=<id,motion,absor ption,transmission,mobility behaviour> (3.1) whereidis the unique identifier ofn,motionis a composed attribute and encloses the mobility attributes of n,absor ption andtransmissionare composed attributes which describe the char-acteristics ofnin terms of absorption and transmission of information respectively, and finally mobility behaviourrepresents the mobility routine and patterns ofn. The following paragraphs describe in detail each of the previous attributes composing a mobile node.

For simplicity purposes, we will also use the termnode when referring to mobile nodes along the rest of this thesis unless we explicitly express the contrary.

ID Theidattribute identifies each mobile node in a unique way from other mobile nodes. We denote byI the set of all mobile nodes identifiers. When referring to theidof a mobile node n, we will use the notationid(n).

We denote byN the set of all mobile nodes and we say thatN is well defined if:

∀n₁,n2∈N ,(id(n1) =id(n2))⇔(n1=n₂)

In practice, the identifier of a mobile node could be the MAC-address of the networking inter-face. In case a mobile node possesses several networking interfaces, only one of them should be selected as the provider.

Motion The motion attribute is composed of thelocation,speed,direction, accelerationand motion potential. The first four attributes simply describe the current location, the speed, the direction of movement and the acceleration of a mobile node. The motion potential attribute refers to the capacity a node has to change its current acceleration attribute. For instance, a mobile phone carried by an elderly person that stays most of the time at home will have a low motion potential as the mobile phone will quite rarely accelerate or slow down. On the contrary, a mobile phone carried by a teenager will have a high motion potential as teenagers are usually hanging out with friends in the city by walking, running, sitting down, taking the bus, etc.

Therefore, they accelerate and slow down constantly and so do their mobile phones.

Formally, the motion of a mobile nodenis defined as an n-tuple:

motion=<loc,speed,dir,a,γ> (3.2)

where,

loc∈R³is the geographical location of the mobile noden, speed∈R⁺is the speed of the mobile nodeninm/s,

dir∈R³is the direction vector of the mobile noden,

a∈R³is the acceleration vector of the mobile nodeninm/s², γ∈[0..1]is the motion potential of the mobile noden.

Concerning the motion potentialγ, a value tending to 0 describes a quasi static node like the one carried by an elderly person and a value tending to 1 describes a mobile phone carried for example by a teenager. When referring to the loc,speed or other motion attribute of a mobile noden, we will use the following notationloc(n),speed(n), etc. It is important to note that all the motion attributes are not constant and they can change throughout time.

Absorption The absorption is composed of theabsorption capacityand the absorption rate.

The absorption capacity represents the maximal amount of bytes that a mobile node allocates to host hovering information. While this attribute is directly related to the local storage resources of a node (e.g.. RAM, SD card, etc.), it is not limited to them and it might also take into consideration other storage resources being available through the network (for instance a laptop connected to the Internet would have a huge storage capacity as hovering information could be stored in devices being located all around the world). Furthermore, the absorption capacity can change throughout time. A mobile node might allocate more RAM for hovering information after the execution of an application that was using a considerable amount of memory. Finally, the absorption capacity is independent from the actual storage capacity of a mobile node. A laptop having a hard disk of 250GB might decide to do not allocate more than some kilobytes for hovering information.

Besides the absorption capacity, the absorption rate defines the rate in bytes per second (Bps) at which a mobile node is able to receive data corresponding to hovering information. This as-pect is directly linked to the mobile node’s networking capabilities and current state. Indeed, coming back to the LostKeys scenario and assuming that the mobile nodes communicate using the 802.11b protocol in ad hoc mode, the absorption rate depends on the characteristics and con-figuration of the 802.11b protocol such as the bandwidth configured to 11 or 55 Mbps. Further, the absorption rate also depends on the current state of the communication channel which might be congested due to the traffic produced by other mobile nodes or due to environment conditions (e.g. signal fading caused by the environment obstacles such as trees or buildings).

Formally, the absorption of a mobile nodenis defined as a tuple:

absor ption=<B,αin> (3.3) where,

B∈Z⁺is the absorption capacity of the mobile nodeninbytes, αin∈Z⁺is the absorption rate of the mobile nodeninBps.

When referring to theB andαin attributes of a mobile noden, we will use the following notationB(n) andαin(n). It is important to note that the absorption attributes are not constant and they can change throughout time. The absorption rate αin of a node n is related to the transmission rateαout as we will see in the next paragraph.

Transmission The transmission is composed of thetransmission rateand thecommunication range. The first attribute represents the rate in bytes per second at which a mobile node is able to transmit to other nodes data corresponding to hovering information, while the second attribute defines the geographical range within which the mobile node is able to transmit hovering in-formation. As for the absorption rate, both transmission attributes are intimately related to the networking capabilities and the current state of the communication channel. If we come back to our scenarios and keep the assumption of all nodes using the 802.11b protocol in ad hoc mode, the transmission range of a node might be around 120m outdoors or 40m indoors, and the transmission rate might be limited to 11 or 55 Mbps. Again, interference and fading of the communication channel would decrease these values.

Formally, the transmission of a mobile nodenis defined as a tuple:

transmission=<αout,rcomm> (3.4) where,

αout∈Z⁺is the transmission rate of the mobile nodeninBps,

rcomm∈Z⁺is the communication range of the the mobile nodeninmeters.

When referring to theαoutandrcommof a mobile noden, we will use the following notation αout(n)andr_comm(n). It is important to note that the transmission attributes are not constant and they can change throughout time. The transmission rateαout and the absorption rateαin (see paragraph Absorption above) of a mobile node nare strongly related. If the mobile nodes are communicating through a single wireless channel, the transmission and absorption rate share the bandwidth of the communication channel (assuming there is no other usage of the bandwidth besides receiving or transmitting hoverinfos). However, if separate channels are used, they are limited by the channel on which they rely.

Figure3.2shows three mobile nodes,n₁,n₂andn₃. Mobile nodesn₁andn₂are able to trans-mit hoverinfos to the other mobile nodes as their communication rangercomm(n₁)andrcomm(n₂) cover the other mobile nodes. However, the mobile noden₃is only able to transmit hoverinfos to mobile noden₁but not ton₂because its communication rangercomm(n₃)is not large enough to letn₂be in range.

n2 n3

rcomm(n1)

rcomm(n2)

rcomm(n3)

Figure 3.2: Mobile nodes and communication range

Mobility Behaviour The mobility behaviour of a mobile node is defined as the mobility pat-terns the mobile node follows throughout time. These mobility patpat-terns can express the common locations between which a mobile node moves around and the time it spends on such locations.

The mobility behaviour of mobile nodes is a result of the daily routine of the person or vehicle carrying it. Indeed, most people have a routinary life and these daily routines can be translated into mobility patterns describing the places along which a person moves throughout the day, and the time spent at each place. Depending on the person and the current context, a mobility pattern may last for only some days (business meetings) or for several months or even years when having the same job, the same friends, the same hobbies, etc. Of course, we can also find people not having any daily routine so that no mobility pattern can be associated with them.

As mentioned previously, the daily routine and therefore the mobile patterns are not limited to people but may, among others, also encompass buses or tramways. Depending on the country, public transport follows precise schedule times which define a very precise mobility pattern. The same reasoning applies to cars of people driving to work as part of their daily routine. We can also consider the route computed by a navigation system of a car as a mobility pattern lasting for the duration of the itinerary.

In the LostKeys scenario, assuming that the festival lasts for one week, the person working on the hot dogs kiosk has a well defined daily routine during the festival. She gets up in the morning to go to the festival, stays at the kiosk almost most of the time, makes a lunching break, goes two/three times to toilets, and finally goes back home at the end of the day. Her mobile phone have then a precise mobility behaviour which is translated into a mobility pattern representing the previous routine.

3.2.2 Hovering Information

A hovering informationor shortly called hoverinfois any kind of information such as a text, photo or video among others, which is stored in a geographical area, called its anchor area, instead of being stored in a particular server or database. This means that hovering information is not linked to a particular storage device as it is the case of most of the information we daily use (e.g. mails, websites, etc. are linked to databases or servers) but to a precise geographical area, its anchor area. As a result, a hoverinfo must rely on the “ephemeral” storage means provided by mobile nodes present in its anchor area. We say ephemeral because mobiles nodes move, fail, run out of battery, or they are simply turned off, making the environment on which a hoverinfo must exist highly dynamic. More precisely, a hoverinfo has to exploit the computing, storage and networking resources of the mobile nodes currently populating its anchor area in order to stay persistent (i.e. stored at its anchor area) and accessible (i.e. ready to be retrieved and used by a mobile application when required). In other words, a hoverinfo has to continuously

Dans le document Hovering information: a self-organising, infrastructure-free information storage and retrieval service for mobile applications (Page 58-74)