Ubi. Com Book Slides (Abridged Version) Chapter 12 Management of Smart Devices, Environments & Interaction Stefan Poslad http: //www. eecs. qmul. ac. uk/people/stefan/ubicom Ubiquitous computing: smart devices, environments and interaction 1
Chapter 12: Overview Chapter 12 focuses on: • Managing Smart Devices in – Virtual ICT Environments – Human User-Centred Environments – Physical Environments • Internal system properties: distributed, autonomous Ubiquitous computing: smart devices, environments and interaction 2
Related Chapter Links • Manage devices as part of human centred activities (Section 5. 6. 5) • OS management of system ICT resources (Section 3. 4) • Intelligent Interaction Management (Chapter 9) • Self, Autonomous System Management (Chapter 8) • Challenges & outlook is related to Management (Chapter 13) Ubiquitous computing: smart devices, environments and interaction 3
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 4
Process and Application Management by the OS • In high resource smart devices MTOS manages ? ? ? (Section 3. 4. 3). • In energy constrained portable devices, OS also manages power (section 4. 3. 4). • In mobile devices, OS & network support for mobile senders & receivers (section 11. 7. 6). Ubiquitous computing: smart devices, environments and interaction 6
Network-Oriented Management • Network viewpoint of ICT systems defines 2 main components (Section 3. 1. 2) – Computer nodes: system management – Network elements to interlink them: network management • Traditionally, a distinction is made between network management and system management although these two are inherently linked Ubiquitous computing: smart devices, environments and interaction 7
Network-Oriented Management • 2 approaches to manage heterogeneous content and applications with different requirements for jitter, delays and packet loss. 1. Keep management support in core network simple 2. management support in core network Ubiquitous computing: smart devices, environments and interaction 8
FCAPS • Why is good system management needed? • System & network downtime leads to lost revenue, opportunities and productivity. Ubiquitous computing: smart devices, environments and interaction 9
FCAPS • Standardization of common network management functions referred to as FCAP • FCAPS functions defines basic requirements for managing distributed computers and hence Ubi. Com systems. • TMN model for managing Open Systems within a telecommunications network defines four logical layers • FCAPS management functions spread across these logical layers. Ubiquitous computing: smart devices, environments and interaction 10
TMN Model
Implementing FCAPS • Many computer service vendors offer client-server type implementations of FCAPS functions via various APIs, • Can support via – – SNMP. ICMP HTTP Open source implementations Ubiquitous computing: smart devices, environments and interaction 12
Simple Network Management Protocol (SNMP) model Main components of model consist of: • Network elements: things to be managed or not • Agent • proxies • Managers • network management information database MIB) • simple network management protocol (SNMP) Ubiquitous computing: smart devices, environments and interaction 13
SNMP
SNMP use to Manage Ubi. Com Can use EDA to support management events including SNMP events (Section 3. 3. 3. 6). SNMP benefits for Ubi. Com • ? ? SNMP challenges for Ubi. Com • ? ? ? Ubiquitous computing: smart devices, environments and interaction 15
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 16
Monitoring • Monitoring obtains the information required to support management functions. • Typical information includes ? • Monitoring can be configured how? Ubiquitous computing: smart devices, environments and interaction 17
Monitoring • . Monitoring involves three separate processes: – analysis – filtering – auditing. Ubiquitous computing: smart devices, environments and interaction 18
Monitoring • 2 main methods or protocols to monitor networked devices or hosts are to use – ICMP – SNMP Ubiquitous computing: smart devices, environments and interaction 19
Accounting • Track service usage and inform authorities about usage and usage costs • Set limits on resource usage. • Automatic handling when usage exceeds limits. Ubiquitous computing: smart devices, environments and interaction 20
ICMP • Internet Control Message Protocol or ICMP can be used to? • Protocol defines? – • ICMP Cons? Ubiquitous computing: smart devices, environments and interaction 21
Configuration Management Objectives: • Set or modify parameters that control routine operation • Track resources defined by their resource descriptions • Track changes in status of resources such as failures • Manage activation and deactivation of resources Ubiquitous computing: smart devices, environments and interaction 22
Configuration Management • A configuration is used as a specification of the settings that are – Variant: e. g. , ? ? – Invariant : e. g. , ? ? • Configuration management involves 4 main management functions: – ? ? • Drivers for reconfiguration ? Ubiquitous computing: smart devices, environments and interaction 23
Re. Configuration Management Side effects of reconfiguration • Systems need to interrupt their service and reboot themselves for reconfiguration changes to take effect. • Configuration conflicts may also become common in multi domains managed Ubi. Com systems • Users can misconfigure devices Ubiquitous computing: smart devices, environments and interaction 24
Reconfiguration Management Easing Configuration & Avoiding Misconfiguration • Ideally, zero manual configuration of devices by users. • Devices should support automatic remote service discovery • Devices should support automatic device installation Ubiquitous computing: smart devices, environments and interaction 25
Fault Management • A fault is defined as the cause of one or more observed error, or abnormal, events. • Fault Management or Safety management concerns maintaining core ICT service operations. • Fault management overlaps with security management How? Ubiquitous computing: smart devices, environments and interaction 26
Fault Management • Detect the fault type of events that lead to system failure. • Organise and manage fault cascades in which a root fault leads to numerous child faults, generally the child faults should be suppressed. • Report faults to an appropriate authority or manager. • Automatic correction and handling of some faults. Ubiquitous computing: smart devices, environments and interaction 27
Fault Management • Fault management is crucial for maintaining the operation of critical infrastructures through monitoring, detecting, preventing and anticipating anomaly events – E. g. , utility distribution, telecomms, transport, logistics, intelligent HVAC, banking, medicine etc. • Fault or safety management involves: – – – – fault prevention, fault prediction, fault event monitoring, fault detection, fault diagnosis, fault handling fault-tolerance. Ubiquitous computing: smart devices, environments and interaction 28
Fault Management • Faults may be random or non-deterministic • Causes? Ubiquitous computing: smart devices, environments and interaction 29
Performance Management • Collect network statistics using polling or event push • Evaluate performance under normal and degraded conditions • Monitor events that exceed thresholds etc. Ubiquitous computing: smart devices, environments and interaction 30
Performance Management • Sometimes it may not be possible to specify absolute single point boundaries for system • There are several specific ways to manage performance: – Best effort – Qo. S – SLA. Ubiquitous computing: smart devices, environments and interaction 31
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 32
Security management • Security management concerns – assessment of risk of threats which cause some loss of value to system assets – heightened through any system vulnerabilities or weaknesses and developing – maintaining appropriate safeguards or security controls to protect assets against threats • 3 basic types of safeguard: – Detection, – Prevention – Correction • Both detection and correction offer a priori protection • Correction offers a posterior protection. Ubiquitous computing: smart devices, environments and interaction 33
Security Safeguards Ubi. Com System security can be modelled in terms of: • Viewpoints of sets of Safeguards that protect the system Assets (the items of value in a system) • against Threats (actions that actively value of assets) • V-SAT model of security Ubiquitous computing: smart devices, environments and interaction 34
V-SAT Model for Security Management 35
V-SAT Model for Security Management Example • Threat – Sender masquerade • Asset – Information about real sender shared with fake sender – Actions requested by a fake sender are performed by receiver. • Safeguards – Authenticate caller identity – Call back real sender – Strong password based access control Ubiquitous computing: smart devices, environments and interaction 36
Security Policies • Security management involves specifying 2 main relationships: – of threats against assets – of safeguards against assets. • A security policy specifies? –. • Security policies are often represented using? Ubiquitous computing: smart devices, environments and interaction 37
Risk Assessment • Risk assessment is used to model: – – the assets of value in a system their loss in value in relation to the probability of the threat happening and to the probability of the threat succeeding. • Information on which to quantify the risk of threats can be based upon? Ubiquitous computing: smart devices, environments and interaction 38
Encryption support for Confidentiality, Authentication and Authorisation • Core security techniques to safeguard distributed systems use – encryption & decryption, – symmetric or asymmetric keys, – to support confidentiality, authentication and authorisation. • Encryption algorithm or cipher: transforms clear or plain content into encrypted content using an encryption key. • Decryption algorithm: transform encrypted content back into plain content using a corresponding decryption key. Ubiquitous computing: smart devices, environments and interaction 39
Encryption support for Confidentiality, Authentication and Authorisation • In symmetric encryption, same key is used for encryption and decryption – Key needs to be kept secret – a secret key. • Key challenge here is how to distribute secret keys while keeping them confidential in transit. – ? ? Ubiquitous computing: smart devices, environments and interaction 40
Asymmetric Encryption • Also called public key encryption algorithms • Public key is made available in an unrestricted fashion and used for encryption by the sender. • Private key, used to decrypt a message in the receiver & is kept secret by the receiver. • Private key cannot be derived from the public key. • This eases the problem of the sender and receiver having to somehow share the same secret key. • Public key encryption enables the public keys used for encryption to be made public. Ubiquitous computing: smart devices, environments and interaction 41
Authentication • Authorisation often involves authentication – • Although public key encryption can be used for authentication, a key issue is how does someone know that the public key belongs to a particular identity? • Anyone could claim that they hold a particular identity? • Something or someone is needed to attest that a particular identity is bound to a particular identity. – ? ? ? Ubiquitous computing: smart devices, environments and interaction 42
Authentication • Mainstream approach involve identity certificate authorities. • In open environment, greater flexibility is useful in order for one party to authorise another party to act on their behalf. • Can some restrictions be removed to enable flexibility? • Cam we have Security without identification: privacy? Ubiquitous computing: smart devices, environments and interaction 43
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 44
Part ? : Overview • An analysis of the use of smart mobile devices reveals an increased risks compared to desktop computers • to secure middleware, • to secure access devices • to secure content, Ubiquitous computing: smart devices, environments and interaction 45
Securing the System and its Middleware Some examples of threats through the use of seamless (wireless) networks: • A) Compromised phones can as free-loader users in a local network; • B) Remote users can overload a network, preventing access by a local user; • C) Local and remote users can eavesdrop on a normal user. Ubiquitous computing: smart devices, environments and interaction 46
Securing the System and its Middleware
Disappearing Security Perimeter Causes • use of seamless networks of IP networks everywhere, • Mobile users • Wireless access: interfere RF interference, eavesdropping • use of Unknown, Untrusted intermediaries: man- in-themiddle attack • connection in a session by a user to the wrong service • These increase risk of security breaches • Solutions? Ubiquitous computing: smart devices, environments and interaction 48
Securing Middleware • Some preventive and corrective type safeguards to handle denial of service threats to mobile devices have also been dealt with elsewhere • Remote Back-ups & occasional synchronisation (Section 12) • Remote-access model can be used to support a distributed UVE type desktop on the mobile device (Section 4. 2) • Mobile devices can be designed to deal with the threat of volatile remote service access (Section 3. 3. 3. 9) • Techniques based upon self-healing and self-protecting can also be used (Section 10. 4) Ubiquitous computing: smart devices, environments and interaction 49
Securing Access Devices may be left unsecure: • Because their owners expect that they will remain under their physical control, • When Devices are discarded: – • Often, devices with inbuilt network security are supplied in a wide open access mode without any security Ubiquitous computing: smart devices, environments and interaction 50
Securing Access Devices Risks of unsecure terminal • Tampering Solutions • tamper-proof devices • tamper-resistant devices • tamper-evident devices Ubiquitous computing: smart devices, environments and interaction 51
Securing Access Devices: To the Owner How to decide who owns & controls a smart device? • A policy model to support secure transient association could be used. – E. g. , Resurrecting Duckling security policy Ubiquitous computing: smart devices, environments and interaction 52
Securing Access in Low Resource Devices • Low resource devices may lack the resources compute encryption or decryption within a reasonable amount of time. – Problems? – Solutions? , e. g. , use Guy Fawkes protocol etc • Low resource devices may be prevented from conserving energy by preventing hibernation or sleep modes in device – – Sleep deprivation attack Barrage attack Problems? Solutions? Ubiquitous computing: smart devices, environments and interaction 53
Securing Information • Access to information is simpler to manage if it can be secured in the static sense at the point of access or where it resides using access control systems based upon policy management. • However, the design of highly distributed systems such as P 2 P systems (Section 3. 2. 6), inherently supports decentralized file sharing, allowing peer users a greater degree of autonomous control over their data and resources. • P 2 P system designs can also be designed to make file sharing anonymous so that a receiver may not known from which other peer computer, its information originated from. Ubiquitous computing: smart devices, environments and interaction 54
Securing Information • The aim of DRM is to secure distribution of content by restricting copying or conversion by consumers and owners – to balance between owners making several copies for personal use on several devices – versus someone distributing content for free or selling on content illegally. • Is DRM a suitable model for distributing content in a Ubi. Com environment? Ubiquitous computing: smart devices, environments and interaction 55
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 56
SOA Management • Ubi. Com system functions can be modelled as services (Section 3. 2. 4) & then managed as services. • Increasingly, systems management occurs – at a human social organisational level using policies and SLAs – to complement management at the ICT level of the system in terms of data throughput and transaction rate. Ubiquitous computing: smart devices, environments and interaction 57
Metrics for Evaluating the Use of SOA • Metrics to evaluate use of SOAs in pervasive environments: – – – service density, service availability, service potential, service impact service redundancy. • Two metrics for service composition are proposed: – service composition length – service composition sustainability. Ubiquitous computing: smart devices, environments and interaction 58
Distributed Resource Management & the Grid • What is a Grid (review)? – • GRID Resource Management System (RMS) component supports adaptability, extensibility, and scalability, allowing systems with different administrative policies: – – – to interoperate while preserving site autonomy, to co-allocate resources, to support load-balancing, to set and maintain quality of service to meet computational cost constraints. Ubiquitous computing: smart devices, environments and interaction 59
Grid model for Managing Ubi. Com Resources Challenges (review)? • Ubi. Com system resources vs. Grid computing resources • Managing resources must be considered at 2 different levels: – system level – application level • Management at both levels simultaneously is challenging because they have different performance goals. Ubiquitous computing: smart devices, environments and interaction 60
SLA Management of Services • In some SOA models, services are modelled more specifically as specifications of sets of operations that can be offered as part of a contract or a Service Level Agreement (SLA) between providers and users. • Contract specifies quantifiable operational targets or outcomes for service levels. • The operation of the services can then be monitored to see if any deviations from these targets occur. • Important applications for use of SLA for management? Ubiquitous computing: smart devices, environments and interaction 61
SLAs & Performance Audits • SLAs need performance models of 2 types of behaviour: – load performance – system performance. • Load models the workload applied to the system and incorporates the behaviour of the users, • System models the performance of services as these process user requests. • SLAs used in Ubi. Com systems also require models of how SLAs for individual services can be aggregated when individual services are combined into composite services Ubiquitous computing: smart devices, environments and interaction 62
Policy-based Service Management Core applications for policy-based management? • Security based policy-based management, particularly access control and user privacy management • Network-based management • Service management, mobility and context-aware management. • User level policy management can be used as part of a vision of i. HCI (Section 5) • Autonomic computing (Section 10. 4). Ubiquitous computing: smart devices, environments and interaction 63
Policy-based Service Management • Instead of managing the detailed low level (re)configuration of parts of the system, the user specifies high-level policies for the system using some policy model. • Policies can e represented in a range of representations – Light-weight KRs to heavy-weight KRs (Section 8. 4) • Policy conflicts can arise because multiple policies may be triggered during the same point of an executing process. Ubiquitous computing: smart devices, environments and interaction 64
Policy Conflicts How to resolve policy conflicts? • Use of priorities • More specific rule takes precedence • Analysis of policies to assess merger of policies, • Use of negotiation (Chapter 9) • Use of voting (Chapter 9) Ubiquitous computing: smart devices, environments and interaction 65
Pervasive Work Flow Management for Services • Workflow as a means to compose & orchestrate services is discussed in Section 3. 3. 4. • Can manage pervasive workflows in terms of distributed control and distributed task assignment. • Here each device to be managed is assigned a role for the workflow and needs to have the local resources (fat client) to execute a local workflow engine. • Devices and the services or work offered can be dynamically discovered and scheduled. Ubiquitous computing: smart devices, environments and interaction 66
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 67
Information Management • Information management is central to Ubi. Com system management. • Any aspect of the system that needs to be managed operations can be modelled as information and then managed using information management techniques. • What Information characteristics that need to be managed? – information volume, persistence, integrity, distribution, discovery, namespace and interoperability. • These in turn are affected by information characteristics – E. g. , hardness, richness and structure – Soft & Rich Information can be equated to intelligent system knowledge representations (Section 8. 4)? Ubiquitous computing: smart devices, environments and interaction 68
Information • Data: raw input into data processes • Output the processed data is called information. • In practice these definitions of data & information overlap. • Focus of data management seems to be information storage , in practice is on information retrieval. • Information can be defined as a structure or syntax • Each type of information application & information provider uses different data structures for storage and retrieval. • Semantics & interpretation of information is often implicit & vaaries Ubiquitous computing: smart devices, environments and interaction 69
Rich versus Lean and Soft versus Hard Information • Current ICT system support for managing data focuses on lean, hard data used for specific operational tasks • However, Individual humans often use richer, softer data for daily activities & strategic business & organisational activities Ubiquitous computing: smart devices, environments and interaction 70
Operational Information Management Organisations’ information management focuses on: • Alphanumeric operational data • Multimedia operational data • Analysing data to make decisions about how well different parts of business are operating. • These are represented as lean, hard data, acquired in TPSs & stored in spread-sheets, file systems or RDBMSs. Ubiquitous computing: smart devices, environments and interaction 71
Personal Information Management • In terms of the use of information in daily activities of individuals there is a similar gap between – operational information to support daily tasks – versus information to support long term user activities & goals. • There are several key challenges here – ? ? Ubiquitous computing: smart devices, environments and interaction 72
Managing the Information Explosion • numbers of Ubi. Com applications • range and No. of interactive devices that can sense the analogue physical world and can read and record multichannel, multimedia content • -> requires an ability to leverage and to cope with data explosion – E. g. , in the My e-Director 2012 project, http: //www. elec. qmul. ac. uk/people/stefan/projects/myed. html, – Aim is to make more audio-video recorded information sources available from multiple camera angles – to allow users to direct which camera angles, episodes in live sports events they view Ubiquitous computing: smart devices, environments and interaction 73
Managing the Information Explosion • Several studies have attempted to estimate the amount of information to be managed. • World produces new data at a rate of 2 -3 exabytes / year. • How much data is generated to record a personal memory of visual and audio inputs of an individual throughout a typical lifetime (~80 years)? • Multimedia content is harder to manage compared to alphanumeric text. Why? Ubiquitous computing: smart devices, environments and interaction 74
Managing Lean & Hard Data: Using RDBMSs Why use RDBMSs to manage data? • ? ? Challenges in using RDBMS to manage lean, hard data ? • Move from network-centric and application centric data manage approach to a data centric management approach which is network and application agnostic. Ubiquitous computing: smart devices, environments and interaction 75
Managing Metadata • Metadata, also referred to annotations, is information to self -describe data. • Ideally metadata should be stored with and bound to the data it is about. • Main benefits of the use of metadata are? • Challenges in using metadata? Ubiquitous computing: smart devices, environments and interaction 76
Managing Metadata • Metadata can be classified as: – application specific versus generic – Intensional vs. extensional data vs. internal data schema • Benefit of a metadata model for Ubi. Com systems to separate these concerns – ? ? ? Ubiquitous computing: smart devices, environments and interaction 77
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 78
Service Management Models for Human User & Physical Environments • Dominant service management model used in smart devices is that they use a Remote Service Access Points (RSAP) model. • This is designed to advertise services in static and centralised directories and to bundle the services on offer at multi-service access portals that can be downloaded and installed locally in order to maintain services on the device. • Other service management models exist which differ with respect to: – how services are accessed – how service changes are managed Ubiquitous computing: smart devices, environments and interaction 79
Service Management Models for Human User & Physical Environments Service Management Models • Stand-alone • Services as Appliances • Remote Service Access Points • Service Contract • User Service Pool • Software as a Service (Saa. S) • Self-managing Ubiquitous computing: smart devices, environments and interaction 80
Models that Reduce Complexity for User Management • For some users, the complexity of learning to use, to operate and to manage such systems is overwhelming because users may attempt to install services that are incompatible or too fragile to use with a particular service infrastructure. • Several types of service management model can minimise user management: – put the know-how and ‘intelligence’ to manage services into a generic service infrastructure, e. g. , the Grid model; – use biologically inspired management. – Use service pools – Use appliance model etc – These are dealt with elsewhere in this chapter. Ubiquitous computing: smart devices, environments and interaction 81
User Task and Activity Based Management • Much of what is termed Personal Computers and personal computing is more suited towards office workers that work on single fixed tasks, in a relatively uninterrupted manner, for long periods of time. • In contrast there are other types of worker activity, that are prone to be interrupted, nomadic, of short duration and where multiple user activities are likely to be interleaved and used to achieve multiple user goals. • In user-centred services, users’ context for ICT events and service reconfiguration can be expressed at multiple knowledge viewpoints, – e. g. , using the mental model of different users. HCI support for this is described in Section 5. 5. Knowledge. Ubiquitous computing: smart devices, environments and interaction 82
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 83
Privacy Requirements: Identity Management Identity management concerns: • Hiding the identity of users unless users give consent or their identity has to be legally known to someone • Managing various identity attributes and tokens of an individual person Ubiquitous computing: smart devices, environments and interaction 84
Biometric User Identification • These identify people by a unique biological characteristic: • Biometric systems can be based upon: – Behavioural trait, e. g. , ? ? ? – Physiological trait, e. g. , ? ? ? Ubiquitous computing: smart devices, environments and interaction 85
Biometric User Identification: Applications • Signature based identification is routinely used. Problems? – E. g. , in banking , but its accuracy is variable. • Finger-print identification is less routinely used – E. g. , to detect crime; for security to access devices & resources – Complemented by use of DNA profiling to identify people. • Retinal scans – at airports can be used for access control Ubiquitous computing: smart devices, environments and interaction 86
Biometric Identification: Pros & Cons • Advantages? • Disadvantages? Ubiquitous computing: smart devices, environments and interaction 87
Content-based Feature Extraction & Classification
Biometric Identification: Design Issues • Typically involves processing a biometric print to extract a multi-dimensional set of features – E. g. , face recognition may involve identifying multiple feature dimensions involving head, eyes, mouth and nose and interrelations. • Design Issues? Ubiquitous computing: smart devices, environments and interaction 89
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 90
Privacy Management • Violation of individual privacy is an oft quoted Ubi. Com peril • Privacy: person’s identity and personal information are kept confidential from others. • Partial privacy rather than complete privacy is used to interact in society and in business – Here a person consents and entrust others with their identity and personal information. – E. g. , ecommerce or voting to elect representatives, . • Full privacy in society, is regarded by some as equally perilous as a lack of privacy. Why? • There are many societal issues concerning privacy. Ubiquitous computing: smart devices, environments and interaction 91
Privacy Requirements • Anonymity versus Authentication • Unlinkability • Unobservability • Notifications, rights and consent Ubiquitous computing: smart devices, environments and interaction 92
Lack of Privacy • Causes for privacy invasion? – Personalisation • It becomes easier to identify who people are and what their behaviours are, referenced in time and space. – Why? Ubiquitous computing: smart devices, environments and interaction 93
Privacy: Safeguards • Several potential safeguards to protect personal privacy: – PET & PIT – Trusted Provider-based Platforms & policies – Privacy legislation • Multilateral approach to personal privacy is needed. Why? How? Ubiquitous computing: smart devices, environments and interaction 94
Privacy-Invasive Technologies versus Privacy Enhanced Technologies • Smart devices can be classified into whether or not they: – Invade privacy, Privacy-Invasive Technologies (PIT) or – Enhance privacy, Privacy Enhanced Technologies (PET) (Clarke, 2001). • However, in practice, there is more of a range from strong PET, through weak PET, to weak PIT, to strong PIT. • Idea of strong PET systems is to avoid use of 3 rd party trust. • Ubi. Com device, designed to act as a weak PET – Can in practice end up being used as a weak PIT. Why? Ubiquitous computing: smart devices, environments and interaction 95
Entrusted Regulation of User Privacy to Service Providers • Platform for Privacy Preferences Project, P 3 P, developed by W 3 C, has been recommended as an industry standard, providing a simple, automated way for users to gain more control over the use of personal information on Web sites that they visit. • At its most basic level, P 3 P is a standardized set of multiple -choice questions, covering all the major aspects of a Web site's privacy policies. • Taken together, they present a clear snapshot of how a site handles personal information about its users. • P 3 P has several weaknesses? – ? ? ? Ubiquitous computing: smart devices, environments and interaction 96
Legislative Approaches to Privacy • In a legislative approach to privacy, collectors of personal information are legally bound to provide a suitable means of notice and consent to users. • However, there are differences in legislative approaches to privacy internationally, – EU Data Privacy Directive incorporates a unique opt-in provision to give consent to personal information access by others. – U. S, data collectors presume consent, and require an affirmative opt -out by the user – Opt-in vs. opt out? Ubiquitous computing: smart devices, environments and interaction 97
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 98
Physical Environment Challenges: Management • Manage single dedicated tasks on embedded system using ASOS • Manage dust-sized devices that may get lost, move outside a prescribed management domain • Manage computer device interaction with physical environment • Manage interactions devices sited in unsupervised, shared physical spaces, e. g. , wall mounted devices • Managing devices throughout their whole environment lifecycle including disposal. Ubiquitous computing: smart devices, environments and interaction 99
Context-Awareness Management 2 aspects of context management considered here: • using context-awareness itself to improve management of systems used for physical world activities and for human world activities • the operational management of context awareness throughout its life-cycle (Section 7. 2). Ubiquitous computing: smart devices, environments and interaction 100
Context-aware Management of Physical and Human Activities • Several core applications of context-aware type systems to aid the management of the use of Ubi. Com systems: – Location-aware management of mobile goods and users to improve distribution or to reach a destination (Section 7. 4), – Location-based access control – Service personalisation (Section 5. 7. . 4) – ICT system management such as managing communication based upon the recipient’s ICT context (Section 7. 6). – Context-aware Power Management (CAPM) – Context-aware Access Control Ubiquitous computing: smart devices, environments and interaction 101
Classifying User Activity as a Composite Context Classifying user activity upon a decision tree for individual contexts
Context Management Challenges • Challenges in managing context-awareness is to perform context adaptation when faced with: – uncertainty, ambiguities, contradictions, logical inconsistencies, conflicts etc (Section 7. 2. 8). • Solutions – Data mining techniques – Prediction models – Learning Ubiquitous computing: smart devices, environments and interaction 103
Management of Tiny Devices • Critical part of the design of micro devices is very efficient power management – in terms of renewable, power storage and power consumption for both sensing, processing and data transmission, Section 4. 3. 4 • Complexity of design & verification as No. of available transistors grows – exponentially increasing productivity gap between these two. – How to tackle this? Ubiquitous computing: smart devices, environments and interaction 104
Managing Tiny Devices • Micro and nano devices may be untethered – i. e. , are airborne or can be fluid born. – Can’t control movement, track them instead – It may be quite difficult to remove all instances of tiny devices , to contain them in a physical space etc • Micro and nano devices may be part of a larger system or environment • Embedded micro devices and macro devices often need to be left unattended for long periods, in relatively inaccessible environments – How to protect unattended devices? • If embedded devices face malicious or inadvertent failures – designed to be self-contained, to fail safe. Why? Ubiquitous computing: smart devices, environments and interaction 105
Chapter 12: Overview The slides for this chapter are split into several parts: • Part A: Core Management of Ubi. Com systems & CCI • Part B: Monitoring, Configuration, Fault & Performance Management • Part C: Security Management Basics • Part D: Security Management of Middleware, Terminals & Content • Part E: Service-Oriented Management of CCI • Part F: Information Management of CCI • Part G: Managing HCI • Part H: Privacy Management: Identification • Part I: Privacy Management: Requirements & Safeguards • Part J: Managing HPI Ubiquitous computing: smart devices, environments and interaction 106
Summary Ubiquitous computing: smart devices, environments and interaction 107
Revision For each chapter • See book web-site for chapter summaries, references, resources etc. • Identify new terms & concepts • Apply new terms and concepts: define, use in old and new situations & problems • Debate problems, challenges and solutions • See Chapter exercises on web-site Ubiquitous computing: smart devices, environments and interaction 108
Exercises: Define New Concepts • Service Appliance Ubiquitous computing: smart devices, environments and interaction 109
Exercise: Applying New Concepts Ubiquitous computing: smart devices, environments and interaction 110
Скачать презентацию Ubi Com Book Slides Abridged Version Chapter 12
4309cf158bfb12222997250ae0f7cab9.ppt