Things insight into the development of 1000 and asked the edge of major changes in computing architecture will affect things do?

The rapid development of global smart phone, push the mobile terminal and the "edge computing" development. The interconnection of all things, all things perceived intelligence community, is networking with attendant physical development, and therefore the edge computing system crashed out.

02.jpg

Edge computing is a distributed computing. Processing, it will run the application data and even some of the features and services to achieve, by the network node into the center edge of the network to reduce the multi-level service delivery, reduce the burden on the core network and transmission.

Whether it is a traditional three-tier system of things, or the new four-tier structure, people ignore a problem, as more and a lot of things future access terminal, is bound to cause new problems of Things data processing! Current methods can effectively solve the edge layer is to strengthen the capacity, so I think that edge computing layer will become an integral part of the new network architecture was.

ECC edge computing Union for edge computing, defines four areas: Device Domain (perception and control layer), network domain (connections and network layer), the data fields (storage and service layer), the application domain (and business intelligence layer). The four "level domain" is calculated by calculating object edges. Internet of Things, the edge computing technology breakthrough means that many control handed over to the cloud without having to implement through local device, the process will be calculated at the local edge layer is completed. This will greatly enhance the efficiency, reduce the load on the cloud. Since more close to the user, but also to provide users with faster response, at the edge of the demand-side solutions.

Edge computing core is close to the user's local data source or provide computing, storage and other infrastructure, and provide cloud services and IT services to the edge of the application environment. Calculating an edge criterion is not only different from the one of the important network 5G 3G, 4G, but also supports low latency linking technology, network technology embodied in the form of specific conditions such as high density, having a scene customization, and other characteristics. Compared to centralized cloud deployments, the edge is calculated not only to solve the delay is too long, the amount of traffic aggregation and other issues, while providing better support for real-time and bandwidth-intensive business.

03.jpg

在国外,以思科为代表的网络公司甚至提出了雾计算,思科已经不再成为工业互联网联盟的创立成员,但却集中精力主导OpenFog开放雾联盟。无论是云、雾还是边缘计算,本身只是实现物联网、智能制造等所需要计算技术的一种方法或者模式。严格讲,雾计算和边缘计算本身并没有本质的区别,都是在接近于现场应用端提供的计算。就其本质而言,都是相对于云计算而言的。

看来在未来的计算领域将会是像云像雾又像风,其实还有一个新提法~霾计算,听上去很可怕,不过核心原理都是尽量靠近设备,靠近终端去即时处理数据。边缘计算中的数据仅在源数据设备和边缘设备之间交换,不再全部上传至云计算平台,防范了数据泄露的风险。

据运营商估算,若业务经由部署在接入点的 MEC 完成处理和转发,则时延有望控制在 1ms之内;若业务在接入网的中心处理网元上完成处理和转发,则时延约在 2~5ms 之间;即使是经过边缘数据中心内的 MEC处理,时延也能控制在 10ms之内,对于时延要求高的场景,如自动驾驶,边缘计算更靠近数据源,可快速处理数据、实时做出判断,充分保障乘客安全。由此可见,5G推动社会从人联时代走向物联时代,连接数的大量增长,叠加边缘计算自身优势,将成为5G时代不可或缺的一部分。

同时,由边缘计算带来的算力需求将成为5G时代重要增量部分。边缘计算联盟ECC正在努力推动三种技术的融合,也就是OICT的融合(运营Operational、信息Information、通讯Communication Technology)。

而其计算对象,则主要定义的了四个领域,第一个是设备域的问题,出现的纯粹的IoT设备,跟自动化的I/O采集相比较而言,有不同但也有重叠部分。那些可以直接用于在顶层优化,而并不参与控制本身的数据,是可以直接放在边缘侧完成处理。

第二个是网络域。在传输层面,直接的末端IoT数据、与来自自动化产线的数据,其传输方式、机制、协议都会有不同,因此,这里要解决传输的数据标准问题。当然,在OPCUA架构下可以直接的访问底层自动化数据,但是,对于Web数据的交互而言,这里会存在IT与OT之间的协调问题,尽管有一些领先的自动化企业已经提供了针对Web方式数据传输的机制,但是,大部分现场的数据仍然存在这些问题。

第三是数据域,数据传输后的数据存储、格式等这些数据域需要解决的问题,也包括数据的查询与数据交互的机制和策略问题都是在这个领域里需要考虑的问题。

最后一个,也是最难的应用域,这个可能是最为难以解决的问题,针对这一领域的应用模型尚未有较多的实际应用。

而从产业价值链整合角度而言,ECC提出了CROSS,即在敏捷联接(Connection)的基础上,实现实时业务(Real-time)、数据优化(DataOptimization)、应用智能(Smart)、安全与隐私保护(Security),为用户在网络边缘侧带来价值和机会。

特别值得关注的是,5G的三大业务场景:eMBB(增强移动宽带)、mMTC (海量机器类通信)和uRLLC(超可靠低时延通信)。每个业务场景都有其自身所面临的一些挑战。例如eMBB将对网络带宽产生数百Gbps的超高需求,从而对回传网络造成巨大传输压力,单方面投资扩容汇聚与城域网络将大幅提高单位媒体流传输成本,无法实现投资收益。uRLLC需要端到端1ms级超低时延支撑,仅仅依赖无线与固网物理层与传输层技术进步,无法满足苛刻的时延需求。mMTC将产生海量数据,导致运营管理的巨大挑战,仅仅由云端集中统一监控无法支撑如此复杂的物联系统。而边缘计算恰好可以为这些问题带来解决方案。由此可见边缘计算有成位5G不可或缺的关键手段!

A few days before the move even raised edge computing as a service (MECaaS) concept. To create the leading "Connect + computing" converged infrastructure, providing full-stack open edge of the cloud service capabilities. By building wide coverage, fixed-mobile converged edge data centers, the introduction of lightweight, easy operation and maintenance of the edge of the cloud platform. Play 5G + edge computing integration advantages, forging a number of industry benchmarking applications, providing industry-leading one-stop construction, operation and delivery capabilities. Joint upstream and downstream industry partners to jointly explore cooperation and win-win business model. However, the edge computing / fog calculated to fall, especially in industry, "application" is the most central issue, the so-called convergence of IT and OT, and more emphasis on the target application in the OT side, that operating system to be achieved .

Edge computing, although fog is calculated low latency, but 50mS, 100mS this cycle for 100μS such a "mission control" high-precision machine tools, robots, high-speed graphics printing system, it is still a very big delay the edge computing so-called "real-time", from the perspective of the automation industry point of view - unfortunately, still are classified in the "non-real-time" applications inside.

Future edge computing cloud computing will sooner or later be pulled from the center of the edge layer, technology evolution will produce enormous economic benefits, not just network vendors, for most SMEs are likely to generate new things accordingly mode .

Focus on cutting-edge research and technology is our business things vigilant question, Amazon, Microsoft, Google, and so Ali has begun its own layout, edge computing layer will cause great changes that are bound to happen.

06.jpg

Things will be the future of the world, let's meet 19-21 November 2019, similar to 2019 (Third) Global Assembly of things - looking thinker!

------ ------ finished article

Disclaimer: This article is "insight - Internet of Things 1000 asked" series eightieth one, intended to hope that through systematic and professional perspective industry will be the networking industry analysis and share the current status of development and economic potential. IOT things, all things Internet, interconnection of all things.

04.jpg

Author: Wang Zhengwei Chairman of Zhongguancun Industry Alliance Secretary-General of the Conference of the world of things 51CTO blog - Novosti sink and micro authorized the release will move Yuan Shuai

Guess you like

Origin blog.51cto.com/13613143/2415176