0% found this document useful (0 votes)
27 views9 pages

Survey of Mobile Edge Computing For Metaverse

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
27 views9 pages

Survey of Mobile Edge Computing For Metaverse

Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).

Please feel free to contact us for questions or remarks.

A Survey of Mobile Edge Computing for the


Metaverse: Architectures, Applications, and
Challenges
Yitong Wang Jun Zhao
[email protected] [email protected]
School of Computer Science and Engineering School of Computer Science and Engineering
Nanyang Technological University Nanyang Technological University
Singapore Singapore
arXiv:2212.00481v1 [cs.NI] 1 Dec 2022

Abstract—Metaverse is an emerging virtual universe where The increasing number of mobile devices and users has
humans can have real-time interactions and solid social links caused the fact that wireless communication networks con-
like in the physical world, and it opens up a new era of stantly evolved to cope with the strict requirement of high
Internet and interactions. In Metaverse, an immersive and
photorealistic environment promotes social activities, including data rate, latency, and machine-to-machine(M2M) connection
education, meetings, and shopping of digital avatars based on density [2]. 5th generational communication network that
critical technologies, including 3D rendering, extended reality, consists of specific attributes is the best choice to cater for the
digital twins, artificial intelligence, and Blockchain. However, emerging Metaverse system. The advent of 5G boosts the per-
the limitations of computation, storage, and energy resources formance indicators of communication networks by a multiple
restrict the development of Metaverse, and a series of system
issues (e.g., latency, security, and battery-life) continue to arise. of 10 to 100 times, which meets the stringent requirements of
As a result, how to find corresponding measurements to mit- various services and applications. It acts as a bridge for the
igate unsatisfactory influences becomes the focus. Mobile edge transmission of original pose data and compressed image data
computing (MEC) as a distributed computing paradigm offloads generated by VR devices and edge cloud servers, respectively,
computation-intensive tasks to the edge of the network. It brings which perfectly enhances the rendering performance of the
the resources as close as possible to the end devices, addressing
the shortcomings mentioned above. In this paper, we propose a Metaverse.
comprehensive survey of the MEC-based Metaverse. Particular Extended reality (XR) [3] technology incorporates mixed
emphasis is given to the technologies convergence, architectures, reality (MR), augmented reality (AR), virtual reality (VR),
and application scenarios, e.g., BoundlessXR and CloudXR.
Significantly, we introduce the potential future directions for
and any other technology that encompasses the fusion of all
developing Metaverse systems. reality and virtualization. XR devices collect data, including
Index Terms—Metaverse, Edge Computing, Extended Reality, biological data like eye-, hand-, and head-tracking, and the
Augmented Reality, Virtual Reality, Blockchain. accumulation of user data from other social media platforms
[4]. XR applications also provide multi-sensor immersiveness
I. INTRODUCTION and real-time interactions for users. Augmented reality(AR)
[5], as a technology that replicates the sensory perception of
Metaverse, as a universe of ternary-world interactions (i.e., the real world in terms of time and space domains, projects
physical world, human world, and digital world), enables additional augmentations generated by computers upon natural
humans as digital avatars in sub-Metaverses to carry out objects, combining real and virtual worlds to enhance people’s
various social activities in the physical world, such as busi- sensory experience further. In the last few years, both KLM
ness meetings, learning, shopping, and digital assets trading Royal Dutch Airlines and British budget airline EasyJet have
[1]. Metaverse significantly revolutionises human manners allowed passengers to check the size of their suitcases for
of interactions with machines, from conventional 2D web boarding through AR technology. AR is performed based on
browsing to immersive real-time interaction in 3D virtual realistic environments, and application services enrich reality
environments, supported by high fidelity image technology, 3D with augmented components [6]. On the opposite side, VR is
rendering, Blockchain, digital twins, artificial intelligence (AI) a part of MR where its surrounding environment is virtual, and
and sophisticated sensor devices. Recently, Metaverse has been people can fully have immersive experiences in virtual worlds
investigated by lots of commercial and gaming companies. For with appropriate wearable equipment (e.g., VR Helmets or
example, Tencent and Facebook have already invested in chat Glasses) [7]. To sum up, VR provides immersive interactiv-
scenarios of Metaverse; even Facebook renamed itself “Meta” ity in a digital world. AR delivers authentic experiences to
in 2021. customers with digital holograms, images, and videos based

1
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

on real-world objects. MR acts as a transition, providing Metaverse. With these concise introductions, readers can
a distinctive experience between VR and AR [8]. Further keep abreast of the ongoing technological developments
developments in micro-sensors and XR technology are making in Metaverse.
XR equipment, such as helmet-mounted displays (HMDs), • We discuss the technological convergence of MEC and
promising to be the primary endpoint for experiencing the Metaverse and illustrate the role that MEC plays in
Metaverse [9]. Metaverse. Unlike previous papers on MEC-based Meta-
The exponential growth of mobile internet traffic has driven verse, we not only describe the contributions of MEC
a dramatic evolution in computing paradigms. Modern mobile to various performance indicators but also emphasize the
applications have more stringent requirements which cloud application of collaborative architectures of MEC with
computing cannot satisfy, such as ultra-low latency, ultra-high other computing paradigms in Metaverse. Specially, we
throughput, ultra-high stability, and high spectral efficiency. point out that MEC is crucial in solving the latency,
Meanwhile, big data-associated Internet of Things devices privacy, and energy challenges in Metaverse.
causes an immense growth in the total traffic flows. To break • We outline the research directions and challenges to pave
the limitations of cloud computing and ensure better cus- the path toward future research attempts to realize higher
tomer service performance, the emerging computing paradigm performance in MEC-based Metaverse architectures. Fur-
of edge computing, especially mobile edge computing, has thermore, our survey aims to assist researchers in gaining
gained the spotlight in both commercial and academic fields a thorough and in-depth understanding of the MEC-based
nowadays. Metaverse and grasping a holistic view of the research in
Following European Telecommunications Standard Institute this area.
(ETSI), Mobile Edge Computing (MEC) is formally defined as The remainder of this paper is organized as follows. Section II
a new platform with significant computing resources, including introduces the present advanced developments of technologies
IT and Cloud-computing and it gets closer to the subscribers’ in Metaverse, and the scenarios of mobile edge computing
side [10]. Intending to extend 3GPP access scenarios such applied to the XR domain. Section III interprets the concepts,
as Wi-Fi, the original MEC concept has now transitioned to architectures, and features of the Metaverse and identifies
“Multi-access Edge Computing” by ETSI. The remarkable some of the current challenges. An introduction to mobile edge
characteristic of MEC is offloading computation tasks from computing, including its strengths and weaknesses, is pre-
core data centres to distributed edge servers located on the base sented in Section IV. Section V summarises the enhancements
stations. Thus, this typically distributed computing paradigm and facilitations when mobile edge computing is applied to the
significantly reduces the communication latency and provides Metaverse and demonstrates the architectures of the Metaverse
a relatively large amount of computation resources compared based on the collaboration of mobile edge computing with
with conventional cloud computing paradigms. Also, the de- cloud and fog computing. Future research directions are shown
ployment of MEC relies heavily on the virtualized platforms in Section VI. Section VII provides the conclusion of the
[11]. paper.
Based on the evolution and combination of the various
technologies, services, and equipment mentioned above, Meta- II. RELATED WORK
verse technology is rapidly gaining attention and enhancement. We have witnessed many papers in terms of different
However, there is a highly distant path to explore to realise aspects of Metaverse technologies and application scenarios
the Metaverse fully. Firstly, transmitting the surging data flows of mobile edge computing. In this section, we discuss recent
together to cloud servers from end devices is contrary to what Metaverse and MEC research and show the main distinction
is being sought. The reason is that it not only aggravates with our paper. Fernandez et al. [12] review an in-depth
network latency on the system but also causes network conges- survey of Metaverse from three main motivation perspectives,
tion and data loss when reaching cloud services via the core including privacy, governance, and ethics design. Specifically,
network. Furthermore, data leakage and asset authentication they analyze privacy from sensory level, behaviour and com-
of the Metaverse are both obstacles encountered. Data security munications, as well as users and bystanders. Also, further
leads to the loss of all personal information and the forgery challenges of these three factors and a novel modular-based
of avatar identities. As multiple administrators provide the framework are presented in this paper. Xu et al. [13] explore
assets and virtual currencies in Metaverse, how to authenticate the significance of using specific spectrums in Metaverse and
and trade among the diverse assets in sub-Metaverses raises a investigate the vision of Blockchain technology, respectively.
concern. Cheng et al. [14] focus on some existing social VR platforms
The contributions of our survey are as follows: that have the potential to evolve into the architectures of the
• We first introduce the relevant background and basic Metaverse and also conduct a test by looking in depth at the
concepts of Metaverse and present a clear overview network operation and capabilities of two typical platforms.
of the types of architectures, evolutionary and eventual Lee et al. [15] present several innovations in the computation
outcomes in Metaverse. Notably, we provide a nuanced arts of Metaverse, proposing some research agendas about
introduction to indicators and essential technologies of democratizing computational arts, digital privacy and safety

2
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

Fig. 1. The Metaverse architecture performs the interactions among virtual area, human area, and physical area

for Metaverse artists, the identification of proprietary rights of virtual scenarios and interactions between physical
in digital artworks, and the directions of technological devel- and virtual worlds; Emerging Web 3.0 converts “server-
opments. For MEC, Siriwardhana et al. [16] emphasize the centric” networks into “user-centric” networks based
survey of mobile augmented reality (MAR) based on MEC and on decentralized networks; MEC servers perform data
discusses future vital application areas for MAR. Lim et al. processing at the edge of the network (i.e., closer to
[17] introduce a study focused on edge-intelligence Metaverse the user) to mitigate system latency; 5G networks with
in term of virtual city development. A stochastic optimal downlink (DL) speeds of 200Mbps meet the constraints
resource allocation scheme (SORAS) is further proposed byNg including low latency, high throughput, and reliability of
et al. based on stochastic integer programming with the aim the Metaverse.
of optimizing the cost-effectiveness of virtual service suppliers Through the ternary nature of the Metaverse definition,
[18]. However, none of these recent papers focuses on the Fig. 1 shows the architecture of the Metaverse, and the
convergence of the Metaverse and MEC paradigm in terms of relationships between these three worlds can be explained in
unique perspectives such as quality of experience, application detail as follows:
scenarios, future challenges, and so on. 1) Physical Area: The physical area provides the essential
III. METAVERSE infrastructures to support the other two worlds, including com-
putation, cache, transmission, and sensors. Data from around
A. Metaverse Introduction the environment and the human body are collected precisely
The term ’Metaverse’ first emerged in the 20th century in by sensor infrastructures. Transmission infrastructures (e.g.
Neal Stephenson’s science fiction novel named ’Snow Crash’ artificial satellites strictly) ensure the stability and continuity
[19], but did not receive much attention at that time. Two of the entire network connection. Computationally intensive
primary factors have driven the Metaverse into focus currently: tasks are processed by servers located at different positions
• Social factors: The Covid-19 pandemic has led to a dra- via robust computation infrastructures, depending on the type
matic shift in the manners of people living and working of these tasks. Also, cache infrastructures at various locations
via the Internet from outdoors to indoors [20], which (i.e., local cache, edge cache, and cloud cache) can store major
has caused an increased frequency of daily activities tasks and reduce the total latency of the entire task processing.
via the Internet. Additionally, commercial enterprises’ 2) Virtual Area: In the Virtual world, the Metaverse engine
strong focus on VR resulted in the re-emergence of the facilitates and manages large volumes of digital information
Metaverse in the public arena; from both the physical and human worlds to enable large-scale
• Technical factors: XR ecosystem developments (includ- Metaverse services [8].
ing AR, VR, MR) facilitate the further enhancement • Engine Layer: Metaverse engine layer consists of various

3
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

technologies to provide a photorealistic environment, each technology continuously emerge. For example, the Bit-
real-time interactions, non-fungible token (NFT) [21] and coin Blockchain supports a token called Bitcoin, while the
translation, respectively. 3D rendering technology pro- Ethernet-backed tokens include SAND, MANA, AXS, and
vides a virtual environment with diverse spatiotemporal GALA. The interoperability of the Metaverse is guaranteed
dimensions and attributes. Digital twins (DT) is crucial to by the accessibility of connections and transactions between
the behaviours of avatars in Metaverse. Especially various different sub-universes or currencies.
human social activities are identified and performed via Multi-technology: A few technological advancements have
DT technology. AI plays a vital role in guaranteeing contributed to the emergence of the Metaverse. For example,
the reliability of Metaverse. By combining with other 5G technology has improved peak transmission rate, time de-
technologies, many machine learning algorithms are used lay, and reliability to meet the requirement that transmissions
to solve challenging tasks, such as computational resource must be completed in milliseconds or less, relieving users
allocation [22], predictions of human actions, and in- of dizziness caused by high latency (>20ms). Also, because
creased spectrum utilisation. downlink speed is proportional to the rendering resolution,
• Virtual Layer: As illustrated in Fig. 1, the virtual layer higher downlink speeds can make virtual environments more
is composed of the virtual environment, digital avatars, realistic. In summary, multi-technology provides an immer-
virtual services/goods and digital assets, which is sup- sive experience and high-fidelity virtual worlds based on
ported by the critical technologies in the engine layer. augmented reality technology, generates a city twin of the
High-fidelity images are provided with 3D rendering physical society with digital twin technology [23], and creates
and AI support by collecting and processing real-world an economic system based on Blockchain technology [24].
environmental data via XR sensor devices. Digital avatars
are capable of behaving exactly like humans in the real
C. Metaverse Technologies
world based on DT technology.
• Virtual Universe: In contrast to the concept of a unified XR: XR technology covers AR, VR, and MR. Wearable
Metaverse, the whole Metaverse comprises a series of VR devices provide users with a completely virtual scenario
dispersed sub-Metaverses where avatars can be served in which they can be fully immersive [25]. For example,
distinctly. Distinct sub-Metaverses can be freed from Second life provides a virtual world in which players can
the constraints of time and space. Due to the existence survive. In this world, players can control their avatars to
of multiple operators, creating multiple sub-universes perform activities and enjoy personalized services. For AR/MR
becomes a fact, but their interconnection holds the key to technology, augmentations are overlaid in the real world [26],
future developments. e.g., users can play Pokémon GO on their mobile phones and
3) Human Area: In reality, people’s intricate social rela- catch virtual Pokemon. XR devices significantly enhance the
tionships and diverse activities contribute to the functioning user’s immersion in the Metaverse.
of society. Crucially, as in the real world, people as avatars Digital Twin: Digital twin (DT) projects humans, objects,
are at the centre of the virtual world and are constantly and environments of the real world into the virtual world,
creating virtual objects. Currently, by wearing devices (e.g. XR creating digital clones that are visually indistinguishable from
HMDs, Wristband sensors), social activities such as business the real world in real-time [27]. By processing the input
meetings, office learning and entertainment concerts can be data, DT is able to manage and optimize physical objects
realised in the virtual world via human-computer interaction periodically [28] [29]. Significantly, the data flows between
(HCI) and especially brain-computer interaction (BCI). Human virtual objects and the physical world are bidirectional. The
activities can be accurately mapped onto the virtual world physical object transmits the collected data in multiple formats
through various Metaverse engines. And services in the virtual to the virtual twins, and the virtual twins convey the processed
world can also continuously influence human interactions and feedback to the physical object [30], further accelerating the
perceptions in the real world. intersection of the human and virtual worlds.
Blockchain: Blockchain is considered to be a decentral-
B. Metaverse Characteristics ized technology [31]. Unlike traditional centralized systems,
Immersive: The ultimate goal of Metaverse development Blockchain technology can help safeguard users’ data, identity
is to approximate the user’s experience in the real world information, and virtual assets with certificates (i.e., NFTs)
gradually. Supported by various infrastructures and Metaverse from leakage and theft in the event of a threat. At the same
engines, users can interact in a rendered and high-fidelity time, Blockchain, as a distributed ledger, ensures that all
virtual universe. Immersive means that the user can adapt to data is protected from tampering and modification. Hence,
the digital world and move around in the Metaverse without Blockchain not only ensures data security and quality but also
discomforts, such as dizziness and nausea. enables seamless data sharing and data interoperability and
Interoperability: Multiple operating systems, such as integrity [32]. Furthermore, the convergence of Blockchain
Ethereum, Solana, and Polygon, co-exist in the Metaverse. and other key technologies can significantly empower the
Accordingly, a variety of value tokens that are derived from performence of the Metaverse system [33].

4
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

AI: AI plays an indispensable role in creating and rendering


large-scale Metaverse. Conventional AI technologies are com-
posed of supervised learning, semi-supervised learning, unsu-
pervised learning, as well as reinforcement learning (RL) [34].
Nowadays, many advanced machine learning (ML) algorithms
of supervised and reinforcement learning have been used
for different challenging tasks, including automatic resource
allocation, attack prevention, and network fault detection [35]
[36]. Based on massive Metaverse multimodal input data,
high-quality Metaverse scenes can be created and rendered
by AI via considerable data interference [37]. Also, deep
learning (DL) and ML algorithms facilitate the provision of
personalized services by making good decisions.

IV. MOBILE EDGE COMPUTING


A. MEC Introduction
The motivation for investigating MEC computing paradigms
is tackling the tradeoff between computation and communica-
tion. As shown in Fig. 2, edge servers are deployed on the
edge of networks, which provides computational resources. As Fig. 2. MEC computing paradigm features the low-latency and energy-
a result, computational tasks are processed at the network’s efficiency procession.
edge as proximate to the user as possible. Different from
fog computing and cloud computing, a MEC server is a
node device which means that the decentralization of mobile act as an edge server at the edge of the Internet of Things
communication networks is achieved [38]. The security and (IoT) system to provide computing resources [42] and various
privacy of networks are enhanced greatly, and congestion services. Meanwhile, the collaboration between vehicle edge
during peak transmission could also be tackled. servers (VES) and fixed edge servers (FES) can provide a
Nowadays, an increasing number of papers on mobile edge variety of offloads and computing options for network com-
computing are investigated, especially in the fields of system puting, offering various edge computing resources depending
architecture design and task offloading. Paper [39] proposes on the tasks. Ren et al. [43] demonstrate that edge servers
a task offloading strategy based on an improved auction can reduce the core network use by elaborating on the system
algorithm, divided into two main phases: task offloading and framework of web-based services with MEC technology and
task scheduling, considering the limited resources of smart further showing that MEC can address real-world Web AR
devices and the stronger computing and storage capabilities deployments.
of edge servers. Furthermore, this strategy provides the basis
V. METAVERSE WITH MEC
for offloading decisions in the decision phase by considering
the time cost of task execution locally or at the edge as well It is theoretically feasible to perform all computing tasks
as energy consumption to determine whether the computation on the local XR devices or cloud servers, but this contradicts
task that needs to be processed will be offloaded to the edge the desire to prolong battery life, lighten device weight and
of the network. Once the task is offloaded to the edge server, maintain ultra-low latency. The convergence of mobile edge
reducing computational latency and reducing transmission en- computing and Metaverse results in a new generation of
ergy consumption to achieve the global optimum are the main MEC technology [44] [45], which is being used to break the
objectives of the task scheduling phase. Paper [40] considers dependency of devices on centralized cloud servers and to
mobility at the user end, classifies cellular networks in more improve the high real-time performance and security of the
detail and gives more accurate state transfer probabilities. And entire system. Primarily, splitting rendering is generated after
this research models the VM migration process as a Markov this convergence. As graphics are heavily rendered, on-device
decision process and uses a policy iteration algorithm to find processes are augmented by being partitioned to IoT devices
the optimal solution, effectively reducing energy consumption and edge clouds. The graphical rendering of the Metaverse
and latency. system is enhanced by using photonic processing for head
tracking and motion tracking of latency-sensitive devices.
B. MEC Application Furthermore, based on the design of stream processing tech-
MEC is used as a key technology that guarantees the nology, Metaverse systems can also effectively decrease the
efficient realisation of diverse services, especially in the prac- latency and have excellent performance as a high-performance
tical application of task offloading and resource allocation. In distributed stream processing system (DSPS) [46]. More stud-
vehicular edge computing (VEC) networks [41], vehicles can ies are explained as follows.

5
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

Fig. 4. BoundlessXR consists of XR devices, 5G network, and edge-cloud.

Furthermore, since edge servers have more computing ca-


pabilities than local devices(e.g. VR devices, HMDs, and AR
glasses), they can efficiently process computation-intensive
tasks and transfer feedback to mobile devices, which will
Fig. 3. Architecture of Mobile Edge Computing-Based Metaverse reduce latency and thus alleviate user discomfort such as
dizziness caused by latency.
System Side: The two main merits of applying MEC to
A. MEC-Based Architecture Metaverse systems today are enhanced resource utilisation
Conventional cloud-based Metaverse architecture is not and significantly reduced latency. Taking advantage of the
beneficial to implementing virtual functions and services via fact that MEC servers come with their own computational
transmitting massive data such as movements of avatars and resources [50], most Metaverse applications can be processed
physical factors of surroundings to the centralized server. To through these servers. At the same time, the wearability and
mitigate negative influences caused by limitations, innovative small size of most devices lead to limitations on the battery
MEC-based Metaverse architectures have resulted in optimiza- life and computational resources, which is tackled by the
tion and a satisfactory performance: MEC-based Metaverse. Simultaneously, the dynamic resource
1) Mobile Edge Computing-Based Metaverse: In the Meta- allocation [23] and edge intelligence-based technology [29]
verse of this architecture, multiple dynamic edge nodes are further improve the performance of all aspects of the Meta-
incorporated to perform users’ instructions, which significantly verse substantially.
decrease the delay caused by users’ movements [47]. As
C. Application Scenarios
shown in Fig. 3, the private data generated by the customers is
transmitted to the specific edge server. Through the processing On-device processing is an essential part of performing
and delivery of the edge servers, customers enjoy a low- Metaverse applications. In a stand-alone model, on-device
latency service and an immersive Metaverse experience. To processing is responsible for dealing with entire XR processing
cope with the data leakage when transmitting users’ data, assignments. While augmented by the edge cloud, on-device
federated learning(FL) is regarded as a critical technology. processing provides rendering and tracking that are power-
Consequently, parameters rather than private data from similar efficiency, high-performance and latency-sensitive. Nowadays,
local models are uploaded to the FL layer for further training. Boundless XR created by the Qualcomm uses split rendering
2) Fog-Edge-Cloud Metaverse: This distributed architecture to distribute the computation tasks between the edge servers
is hierarchical and tiers different servers to eliminate the Meta- and end devices to deliver a genuinely immersive XR ex-
verse fragmentation and computational bottlenecks.Kechadi perience over 5G. Specifically, the architecture is presented
et al. [48] proposed this hybrid computational architecture in Fig. 4. Firstly, VR devices send 6-degrees-of-freedom
for Metaverse services. In cloud layers, Cloud servers exist head pose data to the edge cloud over 5G. The edge cloud
to simulate the virtual worlds. The edge layer servers are sends the compressed image to the devices after processing
dedicated to processing specific virtual buildings’ computation this data and rendering a new image. Finally, VR devices
tasks and providing virtual environment animation. For fog further perform on-device processing. Based on the 5G and
servers located on the fog layer, servers are deployed to MEC technologies, the motion-to-render-to-photon latency is
perform computation tasks of the virtual home environment less than 20ms, which meets Metaverse services’ requirements.
and users’ movements. NVIDIA’s CloudXR is also a novel solution for streaming XR
data from any openVR/XR application on a distant server,
B. Improvement which can also dynamically adjust network conditions and
User Side: In virtual worlds, accurate predictions and maximize image quality while minimizing effective latency.
assessments of users’ behaviours, habitual preferences, and In addition to innovative applications by commercial com-
movement paths are crucial to providing users with an im- panies, deep research in edge computing is also evolving in
mersive and real-time experience. Based on the attributes of academia. A remarkable FACT algorithm was proposed in the
MEC, stable and reliable seamless services could be offered paper [51] to balance three main factors, including network
to users’ devices, and real-time services are also provided as latency, computation latency, and system accuracy in an edge-
a result of the location of MEC servers [49]. based mobile augmented reality system. Applying this new

6
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

paradigm to the MAR applications significantly addressed the provide continuous service and seamless coverage to mobile
optimization problems as a result of multiple mobile users users. In future research directions, mobility management
and improved the performance of processing video frames in can be combined with artificial intelligence. Through this
Metaverse. In order to optimize the computation resources and technologies convergence, the Metaverse system can learn the
AR configurations on the edge servers, another optimization user’s behaviours on its own, predict how the user will move,
protocol called DARE was presented in the paper [52]. When and transmit information about the user’s state to the SDN
considering multiple coherence times in the research of MEC- controller in advance, thus providing a more reliable service.
based MAR, the algorithm based on Lyapunov optimization
in the paper [53] can effectively improve the reliability and C. User Experience
the quality of augments (QoA) in Metaverse systems. Further- Digital avatars experience various activities in photorealistic
more, when studied from the perspective of energy consump- sub-Metaverses and can obtain perceptions through sensor
tion and the interactions between MAR system parameters, devices. The final evolution of the Metaverse is to integrate
the LEAF algorithm was studied by Wang et al. [54] in the with the physical world, which means dark real-world forces
specific energy-aware MEC-based MAR system. also exist in virtual life. For example, robberies and car
accidents in Metaverse badly influence humans in the real
VI. FUTURE RESEARCH CHALLENGES
world and even cause physical pain. Moreover, the unfet-
In this section, future challenges of MEC-based Metaverse tered behaviour of avatars, like sexual assault, also causes
are discussed from the following aspects. psychological victimization of the user. Thus, establishing
A. Computer Resource Allocation a code of ethics to constrain the behaviour of avatars is a
crucial challenge to maintaining a safe community for users in
Currently, papers are primarily conducted on computation
virtual worlds [61]. Furthermore, user addiction [62] is another
allocation. Tang et al. [55] proposed an optimization algorithm
harmful influence on emotional health in Metaverse.
by considering the computation offloading strategies and re-
source allocation problems together and effectively solved the D. Data
interuser interference due to load. Nath et al. [56] studied the
cached content by four decision parameters were considered, For obstacles of Metaverse in the future, two main factors
including whether a given task needs to be cached, how much are supposed to be noticed: data leakage and data reliability.
transmission power is appropriate to utilize during offloading, 1) data leakage: MEC-based metaverse architecture is not
and how many MEC resources to assign to perform a task the same as the decentralized bitcoin architecture. Private
to further improve collaboration issues between MEC servers. and safety-sensitive data (such as avatar ID numbers and
However, these researches do not consider the dynamic na- passwords, digital assets, and currencies) are transmitted to
ture of Metaverse systems. Specifically, the computational the edge cloud from end devices, which may suffer grievous
resources of each node are dynamic in real time rather than attacks by potential adversaries. Lee et al. [63] identify three
static. Therefore, idle computational resources will be unused main ad fraud threats, including blind-spot tracking, gaze and
inefficiently when the MEC is processing computationally controller cursor-jacking, and the misuse of assisted display
intensive tasks after nodes are pre-determined. For further in content sharing. 2) data quality:The data quality levels
research, how to solve the problem of dynamic allocation of directly alter the metaverse applications’ QoS. High data
computational resources is a crucial challenge to be tackled. quality significantly improves the QoE of users, while low
level stands on the opposite side. 3) data reliability: Data
B. Mobility Management sources and generation processes will cause the data to be
Metaverse applications rely heavily on users’ states and unreliable if the sources are fake IDs and jitters occur during
previous behaviours, meaning mobility is a key consideration propagation. Hence, preventing leakage and improving the
when implementing applications sensitive to continuity and quality and reliability of data in Metaverse are also essential
real-time [57]. On the other hand, user mobility and excellent requirements during the development of realizing Metaverse.
mobility management ensure seamless service to enhance the
user experience and QoS further. The alternative approaches E. Delay
are caching [58] and SDN controllers now. For caching, The main goal of the Metaverse is to deliver the processed
the main challenges are how the popularity of the content data to the XR device in real-time via the downlink and to
to be cached can be accurately predicted and how we can render the virtual spatial environment. For example, the delay
achieve collaborative caching across multiple cells to improve generated by VR devices should not exceed 20ms; otherwise,
web caching performance. [59] For SDN controllers, Shah et it will cause spatial vertigo and vomiting to the user in the
al. [60] propose an SDN-based MEC-enabled 5G vehicular Metaverse. Hence, the partitioning of computational tasks is
networks. They use four software modules developed by a constant focus of researchers [64]. Furthermore, balancing
themselves in the SDN controller to develop the rules of QoS the tradeoff between task splitting and delay is essential. For
further. Furthermore, the SDN controller proposed can federate example, Lightweight and wearable XR devices can be worn
and coordinate the allocation of MEC resources to further easily but with fewer computational resources.

7
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

F. Privacy [8] Y. Wang, Z. Su, N. Zhang, D. Liu, R. Xing, T. H. Luan, and X. Shen,
“A Survey on Metaverse: Fundamentals, Security, and Privacy,” arXiv
Data privacy and user identity are the two significant aspects preprint arXiv:2203.02662, 2022.
of Metaverse security that must be ensured. For data privacy, [9] M. Sugimoto, “Extended Reality (XR: VR/AR/MR), 3D Printing,
users’ virtual assets, facial and retinal data, and transaction Holography, AI, Radiomics, and Online VR Tele-Medicine for Precision
Surgery,” in Surgery and Operating Room Innovation. Springer, 2021,
records in the Metaverse are at risk of leakage, especially pp. 65–70.
during data processing and transmission. Moreover, after one [10] E. M. E. Computing, I. Initiative et al., “Mobile-Edge Computing:
user’s behavioural data and interactions with other avatars have Introductory Technical White Paper,” ETSI: Sophia Antipolis, France,
pp. 1–36, 2014.
been collected and analyzed by hackers [19], a fake avatar will
[11] Y. C. Hu, M. Patel, D. Sabella, N. Sprecher, and V. Young, “Mobile
be made that can imitate the original avatar to carry out social Edge Computing—A Key Technology towards 5G,” ETSI white paper,
activities in the Metaverse and even commit crimes. Bystander vol. 11, no. 11, pp. 1–16, 2015.
privacy in the Metaverse is also at risk of being acquired. Then, [12] C. B. Fernandez and P. Hui, “Life, the Metaverse and Everything:
An Overview of Privacy, Ethics, and Governance in Metaverse,” arXiv
the protection of data and identity becomes the way forward. preprint arXiv:2204.01480, 2022.
[13] H. Xu, Z. Li, Z. Li, X. Zhang, Y. Sun, and L. Zhang, “Metaverse
VII. CONCLUSION Native Communication: A Blockchain and Spectrum Prospective,” arXiv
preprint arXiv:2203.08355, 2022.
In this paper, we have introduced an in-depth survey of [14] R. Cheng, N. Wu, S. Chen, and B. Han, “Will Metaverse be NextG
Metaverse based on mobile edge computing. We have pre- Internet? Vision, Hype, and Reality,” arXiv preprint arXiv:2201.12894,
sented an overview of Metaverse and discussed its funda- 2022.
[15] L.-H. Lee, Z. Lin, R. Hu, Z. Gong, A. Kumar, T. Li, S. Li, and P. Hui,
mentals, characteristics, core technologies, and existing issues. “When Creators Meet the Metaverse: A Survey on Computational Arts,”
Afterwards, fundamental research and application scenarios arXiv preprint arXiv:2111.13486, 2021.
like AR(Web-AR)/VR/MR of mobile edge computing are pre- [16] Y. Siriwardhana, P. Porambage, M. Liyanage, and M. Ylianttila, “A
sented. Significantly, we analyze the convergence of these two Survey on Mobile Augmented Reality with 5G Mobile Edge Computing:
Architectures, Applications, and Technical Aspects,” IEEE Communica-
prominent technologies from aspects including architectures, tions Surveys & Tutorials, vol. 23, no. 2, pp. 1160–1192, 2021.
improved peculiarities, and application scenarios. Furthermore, [17] W. Y. B. Lim, Z. Xiong, D. Niyato, X. Cao, C. Miao, S. Sun, and
we show the future research challenges of mobile edge-based Q. Yang, “Realizing the Metaverse with Edge Intelligence: A Match
Made in Heaven,” arXiv preprint arXiv:2201.01634, 2022.
Metaverse. We anticipate that this survey will shed light on [18] W. C. Ng, W. Y. B. Lim, J. S. Ng, Z. Xiong, D. Niyato, and C. Miao,
the direction of the MEC and Metaverse and spark even more “Unified Resource Allocation Framework for The Edge Intelligence-
impressive research in the future. Enabled Metaverse,” arXiv preprint arXiv:2110.14325, 2021.
[19] R. Di Pietro and S. Cresci, “Metaverse: Security and Privacy Issues,”
ACKNOWLEDGEMENT in 2021 Third IEEE International Conference on Trust, Privacy and
Security in Intelligent Systems and Applications (TPS-ISA). IEEE, 2021,
This research is supported in part by Nanyang Technological pp. 281–288.
[20] A. E. Bates, R. B. Primack, B. S. Biggar, T. J. Bird, M. E. Clinton, R. J.
University Startup Grant; in part by the Singapore Ministry Command, C. Richards, M. Shellard, N. R. Geraldi, V. Vergara et al.,
of Education Academic Research Fund under Grant Tier 1 “Global COVID-19 Lockdown Highlights Humans as Both Threats and
RG97/20, Grant Tier 1 RG24/20 and Grant Tier 2 MOE2019- Custodians of The Environment,” Biological conservation, vol. 263, p.
109175, 2021.
T2-1-176.
[21] R. Garcı́a, A. Cediel, M. Teixidó, and R. Gil, “Semantics and Non-
Fungible Tokens for Copyright Management on the Metaverse and
R EFERENCES Beyond,” arXiv preprint arXiv:2208.14174, 2022.
[22] P. Si, J. Zhao, H. Han, K.-Y. Lam, and Y. Liu, “Resource Allocation and
[1] J. Zhong and Y. Zheng, “Empowering future education: Learning in
Resolution Control in the Metaverse with Mobile Augmented Reality,”
the Edu-Metaverse,” in 2022 International Symposium on Educational
arXiv preprint arXiv:2209.13871, 2022.
Technology (ISET). IEEE, 2022, pp. 292–295.
[2] P. Popovski, K. F. Trillingsgaard, O. Simeone, and G. Durisi, [23] Y. Han, D. Niyato, C. Leung, C. Miao, and D. I. Kim, “A Dynamic
“5G Wireless Network Slicing for eMBB, URLLC, and mMTC: A Resource Allocation Framework for Synchronizing Metaverse with IoT
Communication-Theoretic View,” Ieee Access, vol. 6, pp. 55 765–55 779, Service and Data,” in ICC 2022-IEEE International Conference on
2018. Communications. IEEE, 2022, pp. 1196–1201.
[3] J. Guan, J. Irizawa, and A. Morris, “Extended Reality and Internet of [24] H. Ning, H. Wang, Y. Lin, W. Wang, S. Dhelim, F. Farha,
Things for Hyper-Connected Metaverse Environments,” in 2022 IEEE J. Ding, and M. Daneshmand, “A Survey on Metaverse: the State-
Conference on Virtual Reality and 3D User Interfaces Abstracts and of-the-art, Technologies, Applications, and Challenges,” arXiv preprint
Workshops (VRW). IEEE, 2022, pp. 163–168. arXiv:2111.09673, 2021.
[4] J. Evans, “White Paper-The IEEE Global Initiative on Ethics of Extended [25] M. Xu, D. Niyato, J. Kang, Z. Xiong, C. Miao, and D. I. Kim, “Wireless
Reality (XR) Report–Extended Reality (XR) Ethics in Medicine,” 2022. Edge-Empowered Metaverse: A Learning-Based Incentive Mechanism
[5] N. R. D. Torres, G. A. L. Morteo, and R. G. Cruz-Flores, “Conceptual- for Virtual Reality,” in ICC 2022-IEEE International Conference on
ization of Augmented Digital Ecosystem (ADE) in context of Metaverse Communications. IEEE, 2022, pp. 5220–5225.
and The Augmented Dimension,” in 2022 IEEE Mexican International [26] M. McGill, “The IEEE Global Initiative on Ethics of Extended Reality
Conference on Computer Science (ENC). IEEE, 2022, pp. 1–5. (XR) Report–Extended Reality (XR) and the Erosion of Anonymity
[6] T. Braud, C. B. Fernández, and P. Hui, “Scaling-up AR: University and Privacy,” Extended Reality (XR) and the Erosion of Anonymity and
Campus as a Physical-Digital Metaverse,” in 2022 IEEE Conference on Privacy-White Paper, pp. 1–24, 2021.
Virtual Reality and 3D User Interfaces Abstracts and Workshops (VRW). [27] C. A. Lynch, A. O. Adeyeye, and M. M. Tentzeris, “” Smart” Packaging
IEEE, 2022, pp. 169–175. of Self-Identifying and Localizable mmID for Digital Twinning and
[7] M. Qu, Y. Sun, and Y. Feng, “Digital Media and VR Art Creation for Metaverse Temperature Sensing Applications,” in 2022 IEEE 72nd
Metaverse,” in 2022 2nd Asia Conference on Information Engineering Electronic Components and Technology Conference (ECTC). IEEE,
(ACIE). IEEE, 2022, pp. 48–51. 2022, pp. 87–92.

8
This paper appears in the 2022 IEEE 8th International Conference on Collaboration and Internet Computing (CIC).
Please feel free to contact us for questions or remarks.

[28] M. Aloqaily, O. Bouachir, F. Karray, I. Al Ridhawi, and A. El Saddik, Complete Survey on Technological Singularity, Virtual Ecosystem, and
“Integrating Digital Twin and Advanced Intelligent Technologies to Research Agenda,” arXiv preprint arXiv:2110.05352, 2021.
Realize the Metaverse,” IEEE Consumer Electronics Magazine, 2022. [50] W. C. Ng, W. Y. B. Lim, J. S. Ng, Z. Xiong, D. Niyato, and C. Miao,
[29] D. Van Huynh, S. R. Khosravirad, A. Masaracchia, O. A. Dobre, and “Unified Resource Allocation Framework for The Edge Intelligence-
T. Q. Duong, “Edge Intelligence-based Ultra-Reliable and Low-Latency Enabled Metaverse,” in ICC 2022-IEEE International Conference on
Communications for Digital Twin-enabled Metaverse,” IEEE Wireless Communications. IEEE, 2022, pp. 5214–5219.
Communications Letters, 2022. [51] Q. Liu, S. Huang, J. Opadere, and T. Han, “An Edge Network Orches-
[30] Y. Wu, K. Zhang, and Y. Zhang, “Digital twin networks: a survey,” IEEE trator for Mobile Augmented Reality,” in IEEE INFOCOM 2018-IEEE
Internet of Things Journal, vol. 8, no. 18, pp. 13 789–13 804, 2021. Conference on Computer Communications. IEEE, 2018, pp. 756–764.
[31] Z. Wang, Q. Hu, M. Xu, and H. Jiang, “Blockchain-based Edge Resource [52] Q. Liu and T. Han, “Dare: Dynamic Adaptive Mobile Augmented Reality
Sharing for Metaverse.” with Edge Computing,” in 2018 IEEE 26th International Conference on
[32] T. R. Gadekallu, T. Huynh-The, W. Wang, G. Yenduri, P. Ranaweera, Network Protocols (ICNP). IEEE, 2018, pp. 1–11.
Q.-V. Pham, D. B. da Costa, and M. Liyanage, “Blockchain for the [53] J. Liu and Q. Zhang, “Edge Computing Enabled Mobile Augmented
Metaverse: A Review,” arXiv preprint arXiv:2203.09738, 2022. Reality with Imperfect Channel Knowledge,” in European Wireless
[33] S. Badruddoja, R. Dantu, Y. He, M. Thompson, A. Salau, and K. Upad- 2019; 25th European Wireless Conference. VDE, 2019, pp. 1–6.
hyay, “Trusted AI with Blockchain to Empower Metaverse,” in 2022 [54] H. Wang and J. Xie, “User Preference Based Energy-Aware Mobile
Fourth International Conference on Blockchain Computing and Appli- AR System with Edge Computing,” in IEEE INFOCOM 2020-IEEE
cations (BCCA). IEEE, 2022, pp. 237–244. Conference on Computer Communications. IEEE, 2020, pp. 1379–
[34] F. Lin, W. Ning, and H. Zhai, “ICDVAE: An Effective Self-supervised 1388.
Reinforcement Learning Method for behavior decisions of Non-player [55] L. Tang and H. Hu, “Computation Offloading and Resource Allocation
Characters in Metaverse Games,” in 2022 IEEE 10th Joint International for The Internet of Things in Energy-Constrained MEC-Enabled Het-
Information Technology and Artificial Intelligence Conference (ITAIC), Nets,” IEEE Access, vol. 8, pp. 47 509–47 521, 2020.
vol. 10. IEEE, 2022, pp. 1727–1731. [56] S. Nath and J. Wu, “Deep Reinforcement Learning for Dynamic Com-
putation Offloading and Resource Allocation in Cache-Assisted Mobile
[35] T. Huynh-The, Q.-V. Pham, X.-Q. Pham, T. T. Nguyen, Z. Han, and
Edge Computing Systems,” Intelligent and Converged Networks, vol. 1,
D.-S. Kim, “Artificial Intelligence for the Metaverse: A Survey,” arXiv
no. 2, pp. 181–198, 2020.
preprint arXiv:2202.10336, 2022.
[57] Z. Huang and V. Friderikos, “Mobility Aware Optimization in the
[36] X. Zhou, C. Liu, and J. Zhao, “Resource allocation of federated learn-
Metaverse,” arXiv preprint arXiv:2209.08899, 2022.
ing for the metaverse with mobile augmented reality,” arXiv preprint
[58] Y. Cai, J. Llorca, A. M. Tulino, and A. F. Molisch, “Joint Compute-
arXiv:2211.08705, 2022.
Caching-Communication Control for Online Data-Intensive Service De-
[37] O. Bouachir, M. Aloqaily, F. Karray, and A. Elsaddik, “AI-based livery,” arXiv preprint arXiv:2205.01944, 2022.
Blockchain for the Metaverse: Approaches and Challenges,” in 2022 [59] Y. Liu, J. Liu, A. Argyriou, L. Wang, and Z. Xu, “Rendering-Aware VR
Fourth International Conference on Blockchain Computing and Appli- Video Caching over Multi-Cell MEC Networks,” IEEE Transactions on
cations (BCCA). IEEE, 2022, pp. 231–236. Vehicular Technology, vol. 70, no. 3, pp. 2728–2742, 2021.
[38] F. Vhora and J. Gandhi, “A Comprehensive Survey on Mobile Edge [60] S. D. A. Shah, M. A. Gregory, S. Li, R. Fontes, and L. Hou, “SDN-
Computing: Challenges, Tools, Applications,” in 2020 Fourth Interna- Based Service Mobility Management in MEC-Enabled 5G and Beyond
tional Conference on Computing Methodologies and Communication Vehicular Networks,” IEEE Internet of Things Journal, 2022.
(ICCMC). IEEE, 2020, pp. 49–55. [61] J. Yu, A. Alhilal, P. Hui, and D. H. Tsang, “6G Mobile-Edge Empow-
[39] J. Sheng, J. Hu, X. Teng, B. Wang, and X. Pan, “Computation Offloading ered Metaverse: Requirements, Technologies, Challenges and Research
Strategy in Mobile Edge Computing,” Information, vol. 10, no. 6, p. 191, Directions,” arXiv preprint arXiv:2211.04854, 2022.
2019. [62] J. Zhu, P. Ni, and G. Wang, “Activity Minimization of Misinformation
[40] Q. Wang, R. Xie, and T. Huang, “Research on Mobility Management Influence in Online Social Networks,” IEEE Transactions on Computa-
of Mobile Edge Computing,” ZTE TECHNOLOGY JOURNAL, 2018. tional Social Systems, vol. 7, no. 4, pp. 897–906, 2020.
[41] Y. Liu, H. Yu, S. Xie, and Y. Zhang, “Deep Reinforcement Learning [63] H. Lee, J. Lee, D. Kim, S. Jana, I. Shin, and S. Son, “AdCube:
for Offloading and Resource Allocation in Vehicle Edge Computing and WebVR Ad Fraud and Practical Confinement of Third-Party Ads,” in
Networks,” IEEE Transactions on Vehicular Technology, vol. 68, no. 11, 30th USENIX Security Symposium (USENIX Security 21), 2021, pp.
pp. 11 158–11 168, 2019. 2543–2560.
[42] K. Li, Y. Cui, W. Li, T. Lv, X. Yuan, S. Li, W. Ni, M. Simsek, and [64] Y. Cai, J. Llorca, A. M. Tulino, and A. F. Molisch, “Compute-and
F. Dressler, “When Internet of Things Meets Metaverse: Convergence Data-Intensive Networks: The Key to The Metaverse,” arXiv preprint
of Physical and Cyber worlds,” arXiv preprint arXiv:2208.13501, 2022. arXiv:2204.02001, 2022.
[43] P. Ren, X. Qiao, J. Chen, and S. Dustdar, “Mobile Edge Computing–A
Booster for The Practical Provisioning Approach of Web-Based Aug-
mented Reality,” in 2018 IEEE/ACM Symposium on Edge Computing
(SEC). IEEE, 2018, pp. 349–350.
[44] W. Zhang, J. Chen, Y. Zhang, and D. Raychaudhuri, “Towards Efficient
Edge Cloud Augmentation for Virtual Reality MMOGs,” in Proceedings
of the Second ACM/IEEE Symposium on Edge Computing, 2017, pp. 1–
14.
[45] S. Dhelim, T. Kechadi, L. Chen, N. Aung, H. Ning, and L. Atzori,
“Edge-enabled Metaverse: The Convergence of Metaverse and Mobile
Edge Computing,” arXiv preprint arXiv:2205.02764, 2022.
[46] J. Heo, K. Bhardwaj, and A. Gavrilovska, “Poster: Enabling Flexible
Edge-Assisted XR,” in 2021 IEEE/ACM Symposium on Edge Computing
(SEC). IEEE, 2021, pp. 465–467.
[47] L. Chang, Z. Zhang, P. Li, S. Xi, W. Guo, Y. Shen, Z. Xiong,
J. Kang, D. Niyato, X. Qiao et al., “6G-Enabled Edge AI for Metaverse:
Challenges, Methods, and Future Research Directions,” arXiv preprint
arXiv:2204.06192, 2022.
[48] T. Kechadi, L. Chen, N. Aung, H. Ning, L. Atzori et al., “Edge-
enabled Metaverse: The Convergence of Metaverse and Mobile Edge
Computing,” 2022.
[49] L.-H. Lee, T. Braud, P. Zhou, L. Wang, D. Xu, Z. Lin, A. Kumar,
C. Bermejo, and P. Hui, “All One Needs to Know about Metaverse: A

You might also like