Dont Let Me Down Ofﬂoading Robot VFs Up to the Cloud Khasa Gillaniy Jorge Mart ın-Perezy Milan Groshevy Antonio de la Olivay Robert Gazdaz

2025-05-04 0 0 837.15KB 5 页 10玖币

侵权投诉

Don’t Let Me Down!

Ofﬂoading Robot VFs Up to the Cloud

Khasa Gillani∗†, Jorge Mart´

ın-P´

erez†, Milan Groshev†, Antonio de la Oliva†, Robert Gazda‡

Abstract—Recent trends in robotic services propose ofﬂoading

robot functionalities to the Edge to meet the strict latency

requirements of networked robotics. However, the Edge is typi-

cally an expensive resource and sometimes the Cloud is also an

option, thus, decreasing the cost. Following this idea, we propose

Don’t Let Me Down! (DLMD), an algorithm that promotes

ofﬂoading robot functions to the Cloud when possible to minimize

the consumption of Edge resources. Additionally, DLMD takes

the appropriate migration, trafﬁc steering, and radio handover

decisions to meet robotic service requirements as strict latency

constraints. In the paper we formulate the optimization problem

that DLMD aims to solve, compare DLMD performance against

the state of the art, and perform stress tests to assess DLMD

performance in small & large networks. Results show that DLMD

(i) always ﬁnds solutions in less than 30ms; (ii) is optimal in a

local warehousing use case; and (iii) consumes only 5% of the

Edge resources upon network stress.

Index Terms—robotic, optimization, ofﬂoading, Edge

I. INTRODUCTION

NETWORKED robotic services are being adopted to en-

hance operational automation and performance in some

uses cases, e.g., assembly robots in Industry 4.0, or remotely

controlled robots. However, in such use cases, strict latency

requirements [1] are difﬁcult to meet upon network congestion

or large latencies towards the servers hosting the networked

robotic services.

To overcome such limitation, recent works [2], [3] propose

to split the networked robotics functionality into Virtual Func-

tions (VFs), and ofﬂoad them to servers with more computa-

tional resources. Ofﬂoading VFs of a robotic service implies

solving the VF embedding problem. A plethora of works in the

literature have tackled the problem during the last years using

artiﬁcial intelligence [4] and bin packing-alike heuristics [5].

The solutions guarantee that resources – e.g. bandwidth and

CPUs – are not exhausted, and typically minimize the latency

of the embedded service. Recent works have adapted the VF

embedding problem to robotic services – see [2], [3], [6]

– but failed to consider either the latency [2]; radio signal

quality [3], [6]; or robot mobility [3]. Consequently, the robotic

service may not fulﬁll strict service latency requirements,

suffer from low bitrates, or service disruption – due bad radio

connectivity/coverage.

∗Khasa Gillani is with the NETCOM Lab at IMDEA Networks Institute,

†Khasa Gillani, Jorge Mart´

ın-P´

erez, Milan Groshev, Antonio de la Oliva are

@Departamento de Ingenier´

ıa Telem´

atica, Universidad Carlos III de Madrid

‡Robert Gazda is with Future Wireless at Interdigital Inc.

This work has been partly funded by the Spanish Ministry of Economic Af-

fairs and Digital Transformation and the European Union-NextGenerationEU

through the UNICO 5G I+D 6G-EDGEDT and 6G-DATADRIVEN.

far Edge

Cloud

near Edge

service

v1v2

(a)

Cloud

100ms!

(b)

coverage

(c)

Fig. 1: Don’t Let Me Down! fosters ofﬂoading the robot

service VFs up to the cloud (a), yet preventing the cloud large

latencies (b) and running out of coverage (c).

To that end, in this paper we propose Don’t Let Me Down!

(DLMD), an algorithm that fosters ofﬂoading robotic VFs to

the Cloud to minimize the Edge usage while considering (i) the

robotic service latency constraints; (ii) the wireless connectiv-

ity; and (iii) robot mobility. DLMD fosters ofﬂoading the VFs

to the Cloud as long as the latency constraints are fulﬁlled, and

takes VF migration and radio handover decisions to prevent

large latencies and out of coverage situations – see Fig. 1.

II. DON’T LET ME DOWN!PROBLEM FORMULATION

In this section we formulate the VF embedding problem that

DLMD solves to take adequate ofﬂoading, migration and radio

handover decisions for robotic services. We consider a hard-

ware graph Gwhose vertices V(G)correspond to switches

and servers. Speciﬁcally, we consider the three tiers of servers

illustrated in Fig. 1a: Cloud, far Edge, and near Edge; each

with decreasing latency towards the robot, respectively.

The goal of the VF embedding problem is to ofﬂoad robot

VFs minimizing the resource consumption at the Edge, and

satisfying the robotic service constraints. In the following Sec-

tions II-A to II-D we specify the robotic service constraints,

and in Section II-E we formulate the associated VF embedding

problem statement, and prove its NP-hard complexity.

A. Robot computational constraints

The VF embedding problem must ensure that the robot

VFs do not exhaust the computational resources C(n)of each

server n∈V(G):

v∈a(n)

C(v)≤C(n),∀n∈V(G)(1)

That is, the computational requirements of all VFs vassigned

to a computing node a(n) = {v1, v2, . . .}must be lower than

its available computational resources C(n). On top, it must

ensure that all VFs V(Si) = {v1, v2, . . .}of a robotic service

Siare ofﬂoaded at some computing node n:

n∈V(G)

P(v, n)≥1,∀Si∈ S, v ∈V(Si)(2)

arXiv:2210.14208v3 [cs.RO] 14 Feb 2023

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

Don'tLetMeDown!OfoadingRobotVFsUptotheCloudKhasaGillaniy,JorgeMart´n-P´erezy,MilanGroshevy,AntoniodelaOlivay,RobertGazdazAbstractRecenttrendsinroboticservicesproposeofoadingrobotfunctionalitiestotheEdgetomeetthestrictlatencyrequirementsofnetworkedrobotics.However,theEdgeistypi-callyanexpensiver...

展开>> 收起<<

Dont Let Me Down Ofﬂoading Robot VFs Up to the Cloud Khasa Gillaniy Jorge Mart ın-Perezy Milan Groshevy Antonio de la Olivay Robert Gazdaz.pdf

共5页,预览1页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Dont Let Me Down Ofﬂoading Robot VFs Up to the Cloud Khasa Gillaniy Jorge Mart ın-Perezy Milan Groshevy Antonio de la Olivay Robert Gazdaz

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: