| T2T Research Group | J. Hong |
| Internet-Draft | Y-G. Hong |
| Intended status: Informational | ETRI |
| Expires: January 9, 2020 | X. de Foy |
| InterDigital Communications | |
| M. Kovatsch | |
| Huawei Technologies Duesseldorf GmbH | |
| E. Schooler | |
| Intel | |
| D. Kutscher | |
| University of Applied Sciences Emden/Leer | |
| July 08, 2019 |
Problem Statement of IoT integrated with Edge Computing
draft-hong-t2trg-iot-edge-computing-00
This document describes new challenges such as strict latency, uplink cost, uninterrupted services, privacy and security, for IoT services originated from the IoT environmental changes. In order to address those new challenges, the integration of Edge computing and IoT has been emerged as a promising solution. This document discribes the concept of IoT integrated with Edge computing as well as the state-of-the-art of IoT Edge computing. It also proposes an architecture of IoT Edge computing. The direction of Edge computing for IoT should be discussed in the IETF/IRTF.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on January 9, 2020.
Copyright (c) 2019 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Nowadays, most IoT services are based on Cloud computing since it can provide virtually unlimited storage and processing power. The integration of IoT with Cloud computing brings many advantages such as flexibility, efficiency, and ability to store and use data.
However, the IoT environment is changing in such a way that vast amounts of data are created at edge/local networks and about a half of data is stored, processed, analyzed and acted upon close to the data producer. Thus, emerging IoT services introduce new challenges that cannot be addressed by today's centralized Cloud computing models alone.
In this document, we describe new challenges for emerging IoT services such as strict latency, uplink cost, uninterrupted services, privacy and security due to the IoT environmental changes.
In order to address those new challenges for IoT services, the integration of Edge computing with IoT has been emerged as a promising solution. In this document, we describe the concept of IoT integrated with Edge computing as well as the state-of-the-art of IoT Edge computing and propose an architecture of IoT Edge computing. The purpose of this document is to bring up the issues of Edge computing for IoT services in IETF/IRTF.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
Since the phrase 'Internet of Things (IoT)' was coined by Kevin Ashton in 1999 working on Radio-frequency identification (RFID) technology at the Auto-ID Center of the Massachusetts Institute of Technology (MIT) [Ashton], the concept of IoT has been that things connected to the Internet can send and receive information collected by sensors without human intervention, where things are various embedded systems such as home appliances, mobile equipment, wearable devices, etc. IoT has become one of the notable innovations playing an important role in our daily lives [Lin]. IoT is generally characterized by real world small things that are widely distributed but have limited storage and processing power, which involve concerns regarding reliability, performance, security, and privacy.
Cloud computing have been defined in [NIST]: "Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction". Cloud computing has been a predominant technology which has virtually unlimited capacity in terms of storage and processing power. The availability of virtually unlimited storage and processing capabilities at low cost enabled the realization of a new computing model, in which virtualized resources can be leased in an on-demand fashion, being provided as general utilities. Companies like Amazon, Google, Facebook, etc. widely adopted this paradigm for delivering services over the Internet, gaining both economical and technical benefits [Botta].
Now with IoT, we will reach the era of post-Clouds where unprecedented volume and variety of data will be generated by things at edge/local networks and many applications will be deployed on the edge netwoks to consume these IoT data. Some of the applications may need very short response times, some may contain personal data, and others may generate vast amounts of data. Today's Cloud based service models are not suitable for these applications.
It is predicted that by 2019, 45% of the data created in IoT will be stored, processed, analyzed and acted close to, or at the edge of the network and about 50 billion devices will connect to the Internet by 2020 [Evans]. So, moving all data from edge/local networks to the cloud data center may not be an efficient way anymore to process vast amounts of data.
In Cloud computing, users traditionally only consumed IoT data through Cloud services. Now, however, users are also producing IoT data with their mobile devices. This change requires more functionality at edge/local networks [Shi].
Edge computing is a new paradigm in which substantial computing and storage resources are placed at the Internet's edge in close proximity to mobile devices or sensors so that computing happens near data sources [Mahadev]. It works on both downstream data on behalf of cloud services and upstream data on behalf of IoT services. An edge device is any computing or networking resource residing between data sources and cloud-based datacenters. In Edge computing, the end device not only consumes data but also produces data. And at the network edge, devices not only request services and information from the cloud but also handle computing tasks including processing, storage, caching, and load balancing on data sent to and from the cloud [Shi].
The definition of Edge computing from ISO is 'Form of distributed computing in which significant processing and data storage takes place on nodes which are at the edge of the network' [ISO_TR]. And the similar concept of Fog computing from Open Fog Consortium is 'A horizontal, system-level architecture that distributes computing, storage, control and networking functions closer to the users along a cloud-to-thing continuum' [OpenFog]. Based on these definitions, we can summarize a general philosophy of Edge computing as "Distribute the required functions close to users and data".
As the IoT is maturing, systems are converging, deployments are growing, and IoT technology is used with more and more demanding applications such as industrial, automotive, or healthcare. This leads to new challenges for the IoT. In particular, the amount of data created at the edge is expected to be vast. Industrial machines such as laser cutters already produce over 1 terabyte per hour, the same applies for autonomous cars [NVIDIA]. 90% of IoT data is expected to be stored, processed, analyzed, and acted upon close to the source [Kelly], as Cloud Computing models alone cannot address the new challenges [Chiang].
Many industrial control systems, such as manufacturing systems, smart grids, oil and gas systems, etc., often require stringent end-to-end latency between the sensor and control node. While some IoT applications may require latency below a few tens of milliseconds [Weiner], industrial robots and motion control systems have use cases for cycle times in the order of microseconds [_60802]. An important aspect for real-time communications is not only the latency, but also guarantees for jitter. This means control packets need to arrive with as little variation as possible with a strict deadline. Given the best-effort characteristics of the Internet, this challenge is virtually impossible to address with a pure cloud model, when also taking the further challenges into account.
Many IoT deployments are not challenged by a constrained network bandwidth to the cloud. The fifth generation mobile networks (5G) and Wi-Fi 6 both theoretically top out at 10 gigabits per second (i.e., 4.5 terabyte per hour), which enables high-bandwidth uplinks. However, the resulting cost for high-bandwidth connectivity to upload all data to the cloud is unjustifiable and impractical for most IoT applications.
Many IoT devices such as sensors, data collectors, actuators, controllers, etc. have very limited hardware resources and cannot rely solely on their limited resources to meet all their computing and/or storage needs. They require reliable, uninterrupted services to augment their capabilities in order to fulfill their application tasks. This is hard and partly impossible to achieve with cloud services for systems such as vehicles, drones, or oil rigs that have intermittent network connectivity.
When IoT services are deployed at home, personal information can be learned from detected usage data. For example, one can extract information about employment, family status, age, and income by analyzing smart meter data [ENERGY]. Policy makers started to provide frameworks that limit the usage of personal data and put strict requirements on data controllers and processors. However, data stored indefinitely in the cloud also increases the risk of data leakage, for instance, through attacks on rich targets.
Industrial systems are often argued to not have privacy implications, as no personal data is gathered. Yet data from such systems is often highly classified, as one might be able to infer trade secrets such as the setup of production lines. Hence, the owner of these systems are generally reluctant to upload related IoT to the cloud.
As described in section 4, there are new challenges for supporting emerging IoT services and Edge computing is one of the candidates to satisfy these challenges. The motivation for IoT Edge computing was discussed at an Edge computing discussion in IETF/IRTF meetings as follows: [IETF_Edge]
As we described at previous sections, the above motivation for IoT Edge computing could directly be benefits of Edge computing in the IoT environment. The above motivation for IoT Edge computing is mainly related to IoT data and other motivation for IoT Edge computing can exist as other aspects of networking and communication.
In spite of its benefits, Edge computing in IoT services has challenges such as programmability, naming, data abstraction, service management, privacy and security and optimization metrics.
Edge computing can support IoT services independently of Cloud computing. However, Edge computing is increasingly connected to Cloud computing in most IoT systems for processing and storaging data. Thus, the relationship of Edge Computing to Cloud Computing is also another challenge of Edge Computing in IoT [ISO_TR].
As an aspect of IoT, Edge computing can provide many capabilities for IoT services because IoT systems are based on sensors and actuator devices in edge area and IoT data generated from sensors and actuator devices are gathered through a gateway [ISO_TR]. Besides on IoT data, other functions such as computing, control and network functions are also very remarkable to support IoT services. In this document, we will first concentrate on IoT data's aspect since the benefit of Edge computing with IoT data is very big in use cases.
As tremendous IoT sensors, IoT actuators, and IoT devices are connected to the Internet, IoT data volume from these things are expected to increase explosively. And it is expected that much of this high volume of IoT data is produced and/or consumed within edge/local networks, not to traverse through cloud networks. Until now, most IoT data generated by IoT things is transferred and accumulated in a remote server and storage of IoT data in a remote server is expensive in transmission and storage. To mitigate the cost of transmission and storage, it is required to divide IoT data into two types of data; one is stored in edge/local networks and the other is stored in cloud networks. The effect of Edge computing is revealed with the handling IoT data in edge/local networks.
Until now, most network equipment such as routers, gateways, and switches just forward data delivered from other network devices without reading or modifying the content. In end-to-end communication, data is acknowledged and proceed at a final corresponding node. This is a typical usage of cloud computing and a client-server communication. But, in the IoT environment, some IoT data will be transferred to a cloud network and some will be delivered to an edge node. The main reason of this separation is to provide real-time processing and security enhancement in IoT. Although there are many new technologies to reduce the delay and transmission time, it is not easy to guarantee real-time processing. The typical use case of this requirement is industrial Internet and smart factory. Even though there are also several solutions to provide security in IoT, the more basic rule is not to expose the privacy data to public networks. If we separate IoT data into private and non-private data, and keep private data within an edge/local network not to expose them in a public network, the security and privacy in IoT cna be addressed by the separation.
If it is possible to separate IoT data in edge/local networks and cloud networks, Edge computing can do more functions with IoT data in edge/local networks. Because Edge computing has the capabilities to handle IoT data in edge/local networks, it is also possible to analyze IoT data to provide enhanced IoT services such as intelligence. To analyze IoT data in an edge/local network, it is required to have comparatively processing performance and this requirement is not obstacle to deploy Edge computing due to the development of H/W and S/W.
If we consider new challenges of IoT services, not only the big volume of IoT data but also the massive number of IoT things can be a critical problem. Even though, we acknowledge this future problem, the Internet architecture originally has the capability of scalability and it will mitigate scalability issue in the IoT environment. But, we cannot estimate the number of IoT things in the future and we cannot guarantee the Internet architecture still sustain the scalability issue in the IoT environment. Edge computing will separate the scalability domain into edge/local networks and outside network (e.g., cloud networks) and this separation of scalability domain can provide more efficient way to tackle the massive number of IoT things.
Because Edge computing can handle IoT data in an edge area and store the IoT data in an edge node, and proceed IoT data if it is needed, it can also separate the management domain into two parts. Edge Computing can concentrate on management of IoT things in an edge area and cooperate with the management of other outside networks.
When we consider the implementation and deployment of Edge computing, it can be mainly referred to an IoT Gateway. The role of an IoT Gateway is to provide multiple accesses to the heterogeneous IoT devices/sensors, handling IoT data and delivering the IoT data to the final destinations such as cloud networks. Similar to an IoT Gateway, an Edge computing architecture as an edge computing node provides downside connectivity to IoT sensors and devices (southbound connectivity) and upside connectivity to cloud networks (northbound connectivity). Also, the architecture provides the function of data storage. Beside these functions, the Edge computing architecture should provide the computing functions, such as data processing, data analyzing, and additional function of intelligence.
+---------------------------+
| |
| Cloud networks |
| |
+------------+--------------+
|
|
+----------------------+-----------------------+
| | |
| +---------------+---------------+ |
| | | |
| | Edge gateway function | |
| | (Northbound) | |
| | | |
| +---------------+---------------+ |
| | |
| +---------------+---------------+ |
| | | |
| | Edge computing function | |
| | (Storage, Processing, | |
| | Analyzing, Intelligence) | |
| | | |
| +---------------+---------------+ |
| | |
| +---------------+---------------+ |
| | | |
| | Edge networking function | |
| | (Southbound) | |
| | | |
| +-------------------------------+ |
| |
| Edge computing node |
+-----+-------+------+-------+-------+-------+-+
| | | | | |
| | | | | |
+---+----+ | +---+----+ | +---+----+ |
|Sensor 1| | |Sensor 2| .|.. |Sensor n| |
+--------+ | +--------+ | +--------+ |
| | |
| | |
+----+---+ +-----+--+ +-----+--+
|Device 1| |Device 2| .... |Device n|
+--------+ +--------+ +--------+
Figure 1: Architecture of IoT integrated with Edge computing
It is expected that the Edge computing architecture will play an important role to deploy new IoT services with integration to big data and AI services.
This section provides an overview of today's IoT Edge Computing field, based on a limited review of standards, research, open-source and proprietary products in Appendix A. Common aspects of IoT edge computing service platforms are summarized here:
[TBA]
The authors would like to thank Joo-Sang Youn and Akbak Rahman for their valuable comments and suggestions on this document.
| [RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997. |
This list of initiatives, projects and products aim to provide an overview of the IoT Edge Computing. Our goal is to be representative rather than exhaustive. Please help us complete this overview by communicating with us about entries we have missed.
EdgeX Foundry, Home Edge, Edge Virtualization Engine are Linux Foundation projects ([Linux_Foundation_Edge]) aiming to provide a platform for edge computing devices. Such an open source platform can, for example, host proprietary programs currently run on IoT gateway products (Appendix A.2). EdgeX Foundry develops an edge computing framework running on the IoT gateway. Home Edge develops an edge computing framework especially dedicated to home computing devices, controlling home appliances, sensors, etc., and enabling AI applications, especially distributed and parallel machine learning. The Edge Virtualization Engine (EVE) project develops a virtualization platform (for VMs and containers) designed to run outside of the datacenter, in an edge network; EVE is deployed on bare-metal hardware.
This set of open-source projects setup and manage clouds of individual edge computing devices. StarlingX ([StarlingX]) extends OpenStack to provide virtualization platform management for edge clouds, which are distributed (in the range of 100 compute devices), secure and highly available. Akraino Edge Stack, another project from the Linux Fundation Edge [Linux_Foundation_Edge], has a wider scope of developing a management platform adapted for the edge (e.g., covering 1000 plus locations), aiming for zero-touch provisioning, and zero-touch lifecycle management.
Open Edge Computing ([OpenEdgeComputing]) is an initiative from universities, manufacturers, infrastructure providers and operators, enabling efficiently offloading cloudlets (VMs) to the edge. Computing devices are typically powerful, well-connected servers located in mobile networks (e.g. collocated with base stations or aggregation sites). The service platform is built on top of OpenStack++, an extension of OpenStack to support cloudlets. This project is mentioned here as a related project because of its edge computing focus, and potential for some IoT use cases. Nevertheless, its primary use cases are typically non-IoT related, such as offloading processing-intensive applications from a mobile device to the edge.
Multiple products are marketed as IoT gateways (Amazon Greengrass, Microsoft Azure IoT Edge, Google Cloud IoT Core, and gateway solutions from Bosh and Siemens). They are typically composed of a software frameworks that can run on a wide range of IoT gateway hardware devices to provide local support for cloud services, as well as some other local IoT gateway features such as relaying communication and caching content. Remote cloud is both used for management of the IoT gateways, and for hosting customer application components. Some IoT gateway products (Amazon Snowball) have a primary purpose of storing edge data on premises, to enable physically moving this data into the cloud without incurring digital data transfer cost.
Services such as MobileEdgeX provide a platform for application developers to deploy software (e.g. as software containers) on edge networks.
The ETSI MEC industry standardization group develops specifications that enable efficient and seamless integration of applications from vendors, service providers, and 3rd parties across multi-vendor MEC platforms ([ETSI_MEC_03]). Basic principles followed include: leveraging NFV infrastructure; being compliant with 3GPP systems; focusing on orchestration, MEC services, applications and platforms. Phase 1 (2015-2016) focused on basic platform services. Phase 2 (2017-2019) focuses on: supporting non-3GPP radio access technologies, especially WiFi; supporting a distributed, multi-operator and multi-vendor architecture; supporting non-VM based virtualization such as containers and PaaS.
The 3GPP standards organization included edge computing support in 5G [_3GPP.23.501]. Integration of MEC and 5G systems has been studied in ETSI as well [ETSI_MEC_WP_28].
The OpenFog Consortium (now part of the Industrial Internet Consortium) aims to standardize industrial IoT, fog and edge computing. It produced a reference architecture for the Fog ([OpenFog]), which has been published as IEEE standard P1934 in 2018.
The IEEE Fog Computing and Networking Architecture Framework Working Group [IEEE-1934] published the OpenFog architecture as an IEEE document, and plan to do further work on taxonomy, architecture framework, and compliance guidelines.
Named Function Networking ([Sifalakis]) is a research project that aims to extend ICN concepts (especially named data networking) to have the network orchestrate computation. Interests are sent for a combination of function and argument names, instead of using the content name in NDN.
The 5G-CORAL project ([_5G-CORAL]) aims to enable convergence of access across multiple RATs using Fog computing, using for this purpose an Edge and Fog Computing System (EFS).
The FLAME project ([FLAME]) aims to improve performance of interactive media systems while keeping infrastructure costs low. It builds over virtualization technologies such as XOS, OpenStack and ONOS/ODL to offer a programmable media service platform. FLAME leverages IP-over-ICN technology developed through earlier projects including POINT ([POINT]).