缺乏共同标准，物联网还只是个传说

并没有物联网(Internet of Things，IoT)这种东西──关于物联网，是太多的炒作以及纸上谈兵，但它其实还不存在。不过市面上倒是有数种针对不同应用领域的「物连网(Networks of Things)」；这种网络的节点通常不知道该怎么与其他网路上的节点沟通，除非是出产自同样的公司或是由同一个人安装。 我们在iPhone、Android平板设备、Windows平台PC、Macbook或是Linux工作站等不同设备上看到的网页都是一样的，这是因为网际网络有一套共同的内容建立、传递与呈现标准，例如IP、HTTP、HTML，这些共同标准还不存在于今日所谓的「物联网」，却有一些相当完整、但不相容的物联网软件堆迭，包括ANT、 ZigBee与Z-Wave。 根据笔者最近与几位物联网专家的交流，他们的意见都与一位参与两项欧盟赞助大型物联网专案的研究员Levent Gurgen相同：“物联网涵盖了许多应用领域──如智能建筑、家庭自动化、智能城市以及工业自动化系统──其中每个环境都有许多不同的通讯协议，并没有一个共同的沟通方法，这造成一个个不互通的应用系统垂直井，也是物联网的主要障碍。” Gurgen的美国同行们也看到了相同的问题，为此他们成立了一个产业组织，名为工业网际网路(Industrial Internet，工联网)，并获得大厂GE支持；可惜的是，他们打算建立的网络框架结构，仅被视为提升美国竞争力的工具，Levent Gurgen与他欧洲、日本等地的伙伴们也是差不多的打算。这应该只是正常情况，据说中国也有自己国内的物联网计划。 第二页：缺乏黏着剂，物联网还是一盘散沙 第三页：各怀心思，产业结盟费力不讨好

相关阅读：
• 一个高智能化的物联网枢纽--物联网网关
• 让智能照明与物联网擦出绚丽的火花
• 物联网应用中各种无线连接技术对比wcoesmc

{pagination} 缺乏黏着剂，物联网还是一盘散沙 美国柏克莱大学物联网专家、也是物联网新创公司Dust Networks创办人Kris Pister表示：“这个领域被糟糕的标准以及实际安装给拖累了。”好消息是，如同Pister所期望，与网际网络使用的基础标准相同的物联网标准已经呼之欲出。 Pister支持的是6LoWPAN，其最底层是以IEEE 802.15.4 MAC为基础，最上层则为IETF的CoAP协议，其间则是正在成形中的6TSCH ──又称IPv6的时间同步跳频(time-synchronized channel hopping)，也是完成该架构的最后一层必备协议。 此外Pister与同业也正推动一个开放源码软件堆迭OpenWSN，能将6LoWPAN应用于无线感测网路：“我们即将完成第一个完全有效的堆迭，但仍不支援安全性──这是学术机构比较没有兴趣的。”思科等公司也是6LoWPAN的大力支持者，但产业界还是缺乏一个共同协议，这也是物联网成形的“黏着剂”。 举例来说，IBM正在推动的信息序列遥测传输(Message Queuing Telemetry Transport，MQTT)标准，使用的是与6TSCH 不同的事件注册模式(event subscription model)。而就算6LoWAN 的所有支持者达成共识，其他物联网标准的支持者也会基于某些理由自称他们的方案比较好，甚至只是因为历史比较久。 无论是不是每个人都同意，实际情况是他们需要使用已经安装的网络；因此我们还需要一些新的“粘着剂”来让物联网成形，这也是这个令人兴奋的新议题背后最大的问题。 第三页：各怀心思，产业结盟费力不讨好

相关阅读：
• 一个高智能化的物联网枢纽--物联网网关
• 让智能照明与物联网擦出绚丽的火花
• 物联网应用中各种无线连接技术对比wcoesmc

{pagination} 各怀心思，产业结盟费力不讨好 另一家新创物联网公司Millennial Networks 的创办人Sokwoo Rhee则表示，初期对于通讯协议的关注并不是一件好事；该公司所打造的物联网方案是一种表现出色的独家网状网络技术；Rhee指出，发明一堆不同的通讯协议，还不如找对一种应用程序：“我们也在应用程序之前先定协议，结果让我们的进度慢了几年。” 因此Rhee 指出：“从行销的角度，人们会说物联网正兴盛，但以我的观察，该领域还在早期阶段；我跟其他一些同业甚至认为，物联网虽然有在成长，但速度低于预期。”他有机会重新开始，现在正与新兴的工联网联盟合作，期望能为物联网找到“强力粘着剂”。 不过该联盟面临不少障碍，如花了七年时间推动6LoWPAN标准的Pister所言，产业标准订定受限于许多政策；而联盟的成员也将面临挑战，包括 AT&T、GE、IBM与Intel等。顾问公司PTR Group的共同创办人Michael Anderson表示：“这种产业结盟将会消耗大量时间，所产生的效果却有限。” Anderson表示，产业界的观点有太多分歧，而且占据物联网市场其他98%版图的代表都没出声，例如SCADA、制造业、还有汽车产业：“我期待他们也能有机会参与，但是我并不抱太大的希望。” 编译：Judith Cheng 参考原文： Internets of Things Need Glue，by Rick Merritt

相关阅读：
• 一个高智能化的物联网枢纽--物联网网关
• 让智能照明与物联网擦出绚丽的火花
• 物联网应用中各种无线连接技术对比wcoesmc

{pagination} There is no Internet of Things. There's a lot of hype and PowerPoint about the Internet of Things, but it does not exist -- yet. Today a couple dozen Networks (plural) of Things are duking it out in as many market sectors. The typical node on one net doesn't know how to talk to nodes on another net unless it was made by the same company and installed by the same person. You can view this Web page on your iPhone, Android tablet, Windows PC, Macbook, or Linux workstation because the Internet has a common set of standards for creating, transporting, and rendering content such as IP, HTTP, and HTML. These common standards do not exist in the Internet of Things today. Instead there are many fairly complete and incompatible IoT software stacks, from ANT to Zigbee and Z-Wave. I have talked to a half dozen IoT experts in the past week, and they are all saying the same thing as Levent Gurgen, a researcher at CEA-Leti involved in two big IoT projects with backing from the European Union: IoT covers a large number of domains -- smart buildings, home automation, smart cities, and industrial automation systems -- and there are many protocols in each of these environments. There is not one common way of communicating among them. This creates vertical silos of application systems that do not interoperate. This is the main roadblock in IoT today. His colleagues over in the US. see this problem, too. To address it, they are putting together a consortium, calling it the Industrial Internet, a term favored by GE. Unfortunately they see the framework architecture they intend to develop as a competitive advantage for US companies. So much for Levent Gurgen and his partners in Europe and Japan. I suppose it's only fair. I am told China has its own national IoT initiative based in Wuxi. I'd love to get details on it. Here in Silicon Valley, I talked with Kris Pister, a Berkeley IoT researcher and founder of Dust Networks, one of the early IoT startups. He agreed with the assessment of Levent in Europe. "The [IoT] field is held back by lousy standards and lousy implementations," Pister told me. The good news is the industry is close to finishing what Pister hopes is a winning standard based on some of the same underpinnings used in the Internet. Pister backs 6LoWPAN. At its lowest levels it is based on the IEEE 802.15.4 media access controller. At its highest level is the IETF's Constrained Application Protocol. In between, work is about to start on something called 6TSCH (a.k.a. time-synchronized channel hopping for IPv6), one last layer needed to finish the architecture. Separately, Pister and colleagues have been working on OpenWSN, an open-source software stack that implements the 6LoWPAN approach for wireless sensor networks. "We're approaching the point of having the first complete working stack, but we still don't have security -- something universities are less interested in," says Pister. Cisco and others are big backers of 6LowPAN. But there's no universal agreement this will be the glue for IoT. IBM, for example, promotes Message Queuing Telemetry Transport (MQTT), which uses an event subscription model that differs from the approach 6TSCH is taking. Even if the 6LoWAN folks get consensus, no doubt backers of the other couple dozen IoT approaches out there will contend they are better for some reason, even if it is only because they have been around longer. Whether or not anyone agrees, the reality is they may need to work with the networks they have installed. So we still need some new glue to unite the superset of IoTs out there, a sort of IoT super glue. And that's the big problem behind this exciting concept of the Internet of Things. The early focus on protocols was not such a good thing, as it turned out, says Sokwoo Rhee. As founder of Millennial Networks, another one of the early IoT startups, he should know. His company created a whiz-bang, proprietary mesh networking technology for IoT. A better approach would have been to "work on finding the right app rather than inventing a dozen different protocols," he said in an interview with EE Times. "The protocols were defined before the apps, and that put us back a few years." As a result, "from a marketing perspective people would say [IoT is] exploding, but from my perspective of watching it since the early days, I and other entrepreneurs like me would say its growing slower than expected, but it's growing." Rhee has an opportunity for a do-over. He is now one of two presidential fellows working with the emerging Industrial Internet consortium that aims to define the IoT superglue. The consortium faces big hurdles. There's a lot of politics in setting industry standards, says Pister, who has spent seven years working on standards for 6LoWPAN. The group's membership holds challenges, too, says Michael Anderson a co-founder of the PTR Group, a consulting firm. It includes AT&T, Cisco, GE, IBM, and Intel. "Such a marriage will burn lots of hours and produce little substance," Anderson told us. "Their perspectives are too divergent, and there doesn't appear to be representation from the other 98 percent of the [IoT] marketplace. Where are the SCADA folks? Manufacturing? The auto industry?" he asks. "I hope they'll be able to get something going, but I wouldn't hold my breath."

相关阅读：
• 一个高智能化的物联网枢纽--物联网网关
• 让智能照明与物联网擦出绚丽的火花
• 物联网应用中各种无线连接技术对比wcoesmc