Mobile Edge Computing
2020-06-02 11:37:59 0 举报AI智能生成
关于边缘计算的综述的结构图
作者其他创作
大纲/内容
I. INTRODUCTION
brief introduction of it origin etc...
providing a survey of key research progress in this young field
containing an ensemble of promising research directions for MEC.
A. Mobile Computing for 5G: From Clouds to Edges
Mobile Cloud Computing (MCC)
the long propagation
Mission of 5G
support communications, computing, control and content delivery (4C)
require unprecedented high access speed and low latency
information is increasingly generated locally and consumed locally
MEC is implemented based on a virtualized platform that leverages advancements in
network functions virtualization (NFV)
information-centric networks (ICN)
softwaredefined networks (SDN)
A main focus of MEC research
develop these general network technologies
mobile applications
the face recognition
augmented reality (AR)
B. Mobile Edge Computing Versus Mobile Cloud Computing
Low Latency
Mobile Energy Savings
Context-Awareness
Privacy/Security Enhancement
C. Paper Motivation and Outline
a wide-range of issues related to MEC
system and network modeling
optimal control
multiuser resource allocation
implementation
standardization
survey in [36]
potential directions for research and development
content scaling
local connectivity
augmentation
data aggregation and analytics
lacks a survey article
providingcomprehensive and concrete discussions on specific MEC research results with a deep integration of mobile computing and wireless communications
providing relevant surveys on joint radio-and-computational resource allocation for MEC
II. MEC COMPUTATION AND COMMUNICATION MODELS
summarize the basic MEC models, comprising models, based on which the models of MEC latency and energy consumption are developed
A. Computation Task Models
parameters that play critical roles
latency
bandwidth utilization
context awareness
generality
scalability
1) Task Model for Binary Offloading
A highly integrated or relatively simple task cannot be partitioned and has to be executed as a whole either locally at the mobile device or offloaded to the MEC server
2) Task Models for Partial Offloading
be partitioned into two parts with one executed at the mobile device and the other offoaded for edge execution.
affects the procedure of execution and computation offoading
the execution order of functions or routines
some can only be executed locally such as the image display function
task-call graph
can capture the inter-dependency among different computation functions and routines in an application
a directed acyclic graph (DAG)
B. Communication Models
wireless channels differ from the wired counterparts
multipath fading in wireless channels
severe inter-symbol inference (ISI)
a signal being interfered by other signals occupying the same spectrum
interference management becomes one of the most important design issues for wireless communication systems
considerations
joint design of offloading and wireless transmissions
be adaptive to the time-varying channels based on the accurate channel-state information (CSI).
communications are typically between APs and mobile devices with the possibility of direct D2D communications.
D2D communications with neighboring devices provide the opportunity to forward the computation tasks to MEC servers.
different types of commercialized technologies for mobile communications
nearfiled communications (NFC),
radio frequency identification (RFID)
Bluetooth
WiFi
cellular technologies
the key characteristics in Table II
C. Computation Models of Mobile Devices
discuss methodologies of evaluating the computation performance.
CPU performance
the execution latency
energy consumption for local computation
other hardware components
D. Computation Models of MEC Servers
Similar as the mobile devices
The server-computation latency is negligible
consider the nonnegligible server execution time in the general design of MEC systems
Two possible models
deterministic
stochastic
The energy consumption
the usage of the CPU
storage
memory
network interfaces
III. RESOURCE MANAGEMENT IN MEC SYSTEMS
focusing on the research of joint radio-and-computational resource management for different types of MEC systems
A. Single-User MEC Systems
1) Deterministic Task Model with Binary Offloading
the binary offloading decision is on whether a particular task should be offloaded for edge execution or local computation
the problem of transmission-energy minimization under a computation-deadline constraint
minimize the energy consumption for executing a task with a soft real-time requirement
2) Deterministic Task Model with Partial Offloading
3) Stochastic Task Model
by random task arrivals
the long-term average energy consumption and execution latency, are more relevant compared with those of deterministic task arrivals
B. Multiuser MEC Systems
1) Joint Radio-and-Computational Resource Allocation
centralized resource allocation, obtains all the mobile information
distributed resource allocation, using game theory and decomposition techniques
2) MEC Server Scheduling
based on the assumptions of user synchronization and the feasibility of parallel local-and-edge computation
requires relaxation of these assumptions in practical study
3) Multiuser Cooperative Edge Computing
two advantages
burdens on the servers can be lightened
sharing the computational resources among the users can balance the uneven distribution of the computation workloads and computation capabilities over users.
C. MEC Systems with Heterogeneous Servers
1) Server Selection
a key design issue is to determine the destination of computation offloading
2) Server Cooperation
benefits
improve the resource utilization andincrease the revenues of computing service providers
provide more resources for mobile users to enhance their user experience
components
resource allocation
revenue management
service provider cooperation
3) Computation Migration
motivated by the mobility of offloading users
D. Challenges
1) Two-Timescale Resource Management
the task offloading process may across multiple channel blocks, necessitating the twotimescale resource management for MEC
2) Online Task Partitioning
ignore the fluctuation of the wireless channels, and obtain the task partitioning decision before the start of the execution process.
3) Large-Scale Optimization
the increase of the network size renders the resource management a large-scale optimization problem with respect to a large number of offloading decision as well as radio-andcomputational resource allocation variables
IV. AN OUTLOOK FOR MEC RESEARCH
a set of key research directions; analyze the design challenges for each research problem and provide several potential research approaches
A. Deployment of MEC Systems
1) Site Selection for MEC Servers
important factors
site rentals
computation demands
user experience cannot be guaranteed due to the poor signal quality and congestion
obstacles
physical limitations
the computation capabilities are smaller
some of the small-cell BSs may be self-deployed by the home users, and many femto BSs owners may not have the motivation to collaborate with MEC vendors
incur security problems as they are easy-to-reach and vulnerable to external attacks
there exist no available communication infrastructures
need to deploy edge servers with wireless transceivers by properly choosing new locations.
dependent on the computational resource-allocation strategy
2) MEC Network Architecture
design the Het-MEC systems [heterogeneous networks (HetNets)]
the computation capacity provisioning problem is highly challenging and remains unsolved,
exploiting the potential of the service subscriber layer, and utilizing the undedicated computational resources
3) Server Density Planning
to determine the number of edge nodes as well as the optimal combination of different types of MEC servers
challenges
The timescales of computation and wireless channel coherence time may be different
The computation offloading policy will affect the radio resource management policy
The computation demands are normally non-uniformly distributed and clustered
B. Cache-Enabled MEC
1) Service Caching for MEC Resource Allocation
less resources
different mobile services require different resources
two possible approaches
spatial popularity-driven service caching
temporal popularity-driven service caching
2) Data Caching for MEC Data Analytics
should be supported by comprehensive database
imposes extremely heavy burden on the edge server storage
relieved by intelligent data caching that only reserves frequently-used database
how to balance the tradeoff between massive database and finite storage capacity
to establish a practical database popularity distribution model
C. Mobility Management for MEC
challenges for realizing ubiquitous and reliable computing
implemented in the HetNet architecture comprising of multiple macro, small-cell BSs and WiFi APs
users moving among different cells will incur severe interference and pilot contamination
frequent handovers will increase the computation latency and thus deteriorate users’ experience.
1) Mobility-Aware Online Prefetching
issue
Conventional design for mobile computation offloading brings long fetching latency &causes heavy loads
solution
leverage the statistical information of the user trajectory and prefetch parts of future computation data to potential servers during the server-computation time
challenges
the trajectory prediction
the selection of the prefetched computation data
2) Mobility-Aware Offloading Using D2D Communications
handle the user mobility problems in MEC systems
the user mobility brings new design issues
1. how to exploit the advantages of both D2D and cellular communications
2. the selection of surrounding users for offloading should be optimized to account for users’ mobility information, dynamic channels and heterogeneous users’ computation capabilities
3. massive D2D links will introduce severe interference for reliable communications
3) Mobility-Aware Fault-Tolerant MEC
three major and interesting problems
fault prevention
The key design challenges lie in how to balance the tradeoff between QoS (i.e., the failure probability) and energy consumption due to extra offloading links for the singleuser case, and how to allocate protection-clouds for multiuser MEC applications
fault detection
by setting intelligent timing checks or receiving feedbacks for MEC services
fault recovery
can be switched to more reliable backup wireless links with adaptive power control for higher-speed offloading.
4) Mobility-Aware Server Scheduling
this static scheduling design cannot be directly applied for the multiuser MEC systems with mobility due to dynamic environments
D. Green MEC
1) Dynamic Right-Sizing for Energy-Proportional MEC
to switch off/slow down the processing speeds of some edge servers with light computation loads
the profile of computation workload at each edge server should be accurately forecasted
2) Geographical Load Balancing for MEC
leverages the spatial diversities of the workload patterns, temperatures, and electricity prices, to make workload routing decision among different data centers.
MEC servers can coordinate together to serve a mobile user
helps to improve the energy efficiency of the lightly-loaded edge servers as well as user experience
prolong the battery lives of mobile devices
factors
the network congestion state should be monitored and considered
a VM should be migrated/set up in another edge server beforehand, which may cause additional energy consumption
the mutual interests of MEC operators and edge computing service subscribers
the existence of conventional Cloud Computing infrastructures endows the edge servers with an extra option of offloading the latency-critical and computation-intensive tasks to remote cloud data centers
3) Renewable Energy-Powered MEC Systems
Good news
it is reasonable and feasible to power the MEC infrastructures with the state-of-the-art EH techniques.
the mobile devices can also get benefits from using renewable energy
eliminates the need of human intervention such as replacing/recharging the batteries
Problem
the green energy-aware resource allocation and computation offloading
a major concern
The randomness of renewable energy may introduce the offloading unreliability and risks of failure
solutions
can be densely deployed over the system to provide more offloading opportunities for the users.
the chance of energy shortage can be reduced by properly selecting the renewable energy sources
MEC servers can be powered by hybrid energy sources to improve reliability
, wireless power transfer (WPT), which charges mobile devices using RF wave
E. Security and Privacy Issues in MEC
1) Trust and Authentication Mechanisms
,thebasicideaistoknowtheidentityoftheentity that the system is interacting with
the conventional trust and authentication mechanisms designed for Cloud Computing systems inapplicable
there will be a large number of edge servers serving massive mobile devices
2) Networking Security
challenges
the difficulties in the distribution of credentials
techniques such as SDN and NFV are softwares by nature and thus vulnerable
3) Secure and Private Computation
V. STANDARDIZATION EFFORTS AND USE SCENARIOS OF MEC
A. Referenced MEC Server Framework
B. Technical Challenges and Requirements
1) Network Integration
2) Application Portability
3) Security
4) Performance
5) Resilience
6) Operation
7) Regulatory and legal considerations
C. Use Scenarios
1) Video Stream Analysis Service
2) Augmented Reality Service
3) IoT Applications
4) Connected Vehicles
VI. CONCLUSION
It aims at enabling Cloud Computing capabilities and IT services in close proximity to end users, by pushing abundant computation and storage resources towards the network edges.
key components of MEC systems
the computation tasks
communications
mobile devices
MEC servers computation
0 条评论
下一页
