PREDIX - The Industrial Internet Platform	Category: ICT Published: 2016 #1708b
	General Electric Co.	17622u/18206r

Title	Predix - The Industrial Internet Platform	Predix - 産業インターネット基盤
Index	Introduction: Rapid explosion of sensors: Edge-to-Cloud Platform: Industrial-grade Security:: Digital Twin: Growing the Ecosystem:	序: センサーの量的爆発: エッジ-to-クラウドプラットフォーム: 産業グレード・セキュリティ: デジタル･ツイン: エコシステムの成長:
Tag	; Always-on; APM; BLOB; Brownfield-Greenfield; CCM; Cloud; Cloud Foundry; Connectivity; Context-adaptive; CSA-CCM; Data driven; DDS; DevOps; DevOpsSec; Digital Twin; Future-proof; Edge; Heap-map dashboard; HIPAA; IIC; III; ISO27001/002; ITAR; MatLab; MODBUS; Multi-tenancy; MVC; OPC-UA; OT; Pay-as-you-go; PLC; PoPs; Predict & prevent; Predix Edge; Protocol-agnostic; RoR; SDI; SDK; SDM; Silo; SOC; Software platform; Tier Ⅲ, Tier Ⅳ; TimescaleDB; UX; ZTP;

Résumé

Remarks

>Top 0. Introduction: This is the summary of 'Predix', developed by GE, as the Industrial Internet Platform was published in Nov. 2016, which is useful to understand the concept of IoT.; In IoT age; it is believed that the advanced factories are connected by the Internet. Here is a big question mark. Because the factory should be most strictly operated without bug or delay or disconnection. But the Internet is the world of free but unsecured network, with full of usual delay or unreliability. I'd like to know how it is possible to connect the most reliable place with most unreliable network? Or how the most holy Cathedral becomes open and be connected by most crowded Bazar with full of good and evil.

0. 序: 2016/11公表のGEのPredixの概略である。IoT時代には先進的な工場はインターネットと接続される。 ここでの疑問は、最も厳格な運用が求められる工場と信頼性の低いインターネットとがどう接続されるのか。 いわば聖堂の扉を開き、喧噪のバザールとどのように接続するのだろうか。

>Top 1. Rapid Explosion of Sensors: Companies are transforming business models by IIoT (Industrial Internet-of-Things: rapid explosion of sensors, Ultra-low cost connectivity, and data storage together with powerful analytics. >Top Data driven transition in technology: This transition will require new software development and analytics. however, most transitions fail because underestimating the complexity, pervasiveness and organizational impact. Industrial data; less than 3% is tagged and used. Datasets can be fragmented and siloed. >Top OT (Operational technology) and IT systems often operate separately. Edge devices are not always connected, or air-gapped due to security, or need to continue operation when the connection is unavailable. Applications need to adapt to local conditions at the edge in real-time. Best practices are used in some but not all parts of an organization. KPIs are not standardized across the business. How decision makers are struggling to answer questions: to allocate budgets strategically. to perform the facility optimally. to check security policy putting us at risk. to enhance customer experience by using the edge data. The power of an Industrial Internet Platform can help to get the answers: be machine-centric heterogeneous data acquisition, storage, management, access. predictive analysis guide personnel with intuitive user experiences. be delivered securely in the cloud and at the edge. Precox: that's why GE built Predix. Businesses can create innovative apps on Predix that turn real-time operational data into actionable insights. GE also offers discrete SaaS solutions as APM (Asset Performance Management). >Top Why a Software Platform?: is designed a reusable building block approach. Build apps quickly Leverage work elsewhere Reduce sources of error Develop and share best practices Lower risk of cost and time >Top Future-proof their initial investments Independent thierd parties can also build apps, allowing businesses to extend their capabialities. >Top Why in the Cloud? >Top Lower costs based on a centrally managed and shared infrastructure in a pay-as-you-go subscription model. Scale to meet different business, adjusting capacity on-demand. Generate actionable insights with assets. Deliver insights from analytics. Enable improved governance, standardized security, and release management control and consistency. >Top Why at the Edge? Edge and cloud architectures are complementary. Reduce latency for mission-control. Adhere to SLAs and regulatory compliance.. Avoid unnecessary exposure of data. Offload compute-intensive task - such as analytics - from resource constrained devices. Why move Today? to decrease unplanned downtime, increase productivity and minimize missed opportunities. to start with an industrial edge-to-cloud platform today; an extensibl architecture to meet future requirements. Delivering outcomes: to create differentiated services that generate new sources of revenue. Scheduling & Logistics: improving performance & efficiency; can result in lower repair costs. >Top Connected Products: replace the current break-fix model with a predict-&-prevent services by making machines software defined. Intelligent Environments: tap into sensors to collect and analyze data. Field Force Managment: give workers the machine data, epertise, and processes to make repairs and upgrades. Industrial Analytics: monitor asset helth to identify problems. >Top Asset Performance Managment (APM): achieve reliability and availability throughout the life cycle of assets. Operations Optimization: use key insights on an enterprise-wide scale to resolve operational issues.

1. センサーの量的爆発: IoTによる会社の変化 センサーの普及 安価な接続 強力な分析ツール データ･ドリブン技術への移行 今までデータ利用は3%以下 データは個別分散 (孤立したサイロ状態) OTとITは別稼働 エッジデバイスは、ネットワーク断絶の場合も稼働条件 エッジのアプリはローカル条件次第 ベストプラクティスは部分的 KPIsは標準化されていない 一方経営者とっては以下の課題あり: 予算の戦略的配分 工場の最適稼働 セキュリティポリシイの点検 エッジデータ活用による顧客への効果 Industrial Internet Platformによって上記課題が解決する。 マシン中心 雑多なデータ取得･蓄積･管理･アクセス 予測分析 なぜソフトベースで設計･提供するのか: 再利用可能な積み木方式として 迅速に構築可能 他での活用も可能 エラーの減少 ベストプラクティスを目標に開発･共有する 時間･費用コストリスク減少 将来性の確保も考慮 なぜクラウドか: 安価、使用分のみの支払方法 (Pay-as-you-go) 規模の自由度 資産活用の自由度 分析可能 セキュリティ、販売時期管理 なぜエッジか: エッジとクラウドは本来補完的 遅延時間なし SLAと諸規則に拘泥 データの守秘 分析など集中計算力 なぜ今なのか: 機会損失をなくす エッジ・クラウドプラットフォームなので将来の拡張性確保 サービスの差別化による新規収益追求 達成目標: 生産計画・納入 ソフト導入による機械故障予知機能向上による生産連携 センサーによるデータ収集 現場ワーカーを支援 稼働率分析 資産の稼働管理 (APM): 資産のライフサイクルを通じて信頼性･可用性の達成 操業の最適化: 企業視点での操業課題解決 Scope of Connectivity: Apps, Analytics Distributed Data Interoperability & Management Connectivity Framework Connectivity Transport Connectivity Network [Neck] Link Physical

>Top 2. Edge-to-Cloud Platform: Public clouds don't support the unique and demanding requirements of industry: Public clouds are developed to support IT data (ERP or CRM). Industrial data exists as multi-terabyte, sampled continuously. (Eg. One plane flight can generate 1TB.) Public clouds don't integrate edge service and intelligence. Customers need third-party add-ons that are not integrated into the platform. Without an integrated edge data must be centralized; may be impractical from a cost and privacy perspective. Intelligence at the edge also allows the business to meed rigorous performance; with real-time decision-making that can dramatically affect the way assets behave. >Top Predix Edge includes a range of integrated technologies; Predix Machine, Predix Connectivity, and Predix EdgeManager. >Top GE and non-GE devices use edge-based Predix Machine; as a "SDM (software defined machine)." Predix Machine runs on a wide variety of hardware platforms from sensors, controllers, gateways, to on-premise appliances. The software provides security, authentication, and governance services for endpoint devices; ensuring that assets are connected, controlled. Predix Machine provides advanced edge analytics such as in-motion data analytics, machine learning capabilities, and analytics deployment at the edge. Predix connectivity: The design and initial deployment of connectivity services takes 6-12 months. offers seamless, secure and reliable end-to-end communication between gateway and devices, and Predix Cloud, including fixed line, cellular, and satellite. End-to-end connectivity solutions from the edge to Predix: End-to-end route and flow management between edge and cloud. >Top Protocol-agnostic network configuration and management for M2M and M2C (Machine-to-Cloud) connectivity. Centrally management driving QoS and bandwidth optimization Policy-driven data forwarding between between cloud and on-premises. Physical connectivity globally via cellular, fixed or satellite. Secure VPN between the edge and cloud. Manage the edge assets by providing VNC, RDP, SSH, HTTP. End-to-end monitoring One-stop-shop billing and reporting >Top Zero touch provisioning (ZTP) with a self-management portal. Predix EdgeManager; which eases the management configuration and administration of edge devices. can quickly determine device condition and connectivity health. can be auto-enrolled and decommissioned Predix is: >Top Pivotal's Cloud Foundry: support programing tools with DevOps environment, scale applications in hours or days. available in global PoPs (Points-of-Presence): operate at either Tier Ⅲ (concurrently maintainable) or Tier Ⅳ (fault tolerance) levels. Capacity On-Demand: uses software-defined infrastructure (SDI) as an abstraction layer above the hardware, evolving over time. Provision management can be done at a granular level. Enhanced Security Controls: include encryption, key management, incident response services, incident response services, logging, network-level security, support for end-to-end chain of custody reporting, and 24x7 security operations centers. Modeling assets: enables developers to create, store, and manage asset models that define asset properties between assets and other modeling elements. Templates can be used to create the structures the define the components that make up a complex asset. Data capture, processing and management: Rapid access to data and timely analytics while minimizing storage and compute costs. >Top Secure multi-tenancy model that includes network-level data isolation and encrypted key-management Connection to the source; with GE and non-GE machine sensors, etc. Data ingestion; from the source in real time. Pipeline processing; efficiently ingest massive amount of data; to be converted to the correct format. Data management; be stored in the appropriate data store (machine sensor data, BLOB (Binary Large Object), RDBMS; tools to extract value from these data. Turning insights into outcomes. Operational analytics: Data is analyzed in real time at the edge. Historical analytics (petabytes); to build a large scale predictive model. Descriptive analytics: what happened and why. Predictive analytics: what might happen next by forecasting a model. Prescriptive analytics: help improve the decsion-making process; to determine possible actions towards a solution. >Top The Predix User Experience (UX) system: provides developers with simple, modular and cohesive solutions, layers and UI components. applications are not only context-aware, but also context-adaptive; mobile staff, field operators, plant managers, business analysts, and data scientists can visualize data in the right context, built their own data models, answer key questions, and deliver on business outcomes. Mobility for always-on: Predix Mobility provides an software developer kit (SDK) and a rich set of cross-platform responsive components.; this will allow developers to synchronize data between mobile devices and enterprise data domains. Geospatial intelligence for enhanced insights: Predix location and mapping services provides precise, location-based information; can be critical in areas such as field-service, transportation logistics, supply chain inventory management, and risk management. >Top To take advantage of the Industrial Internet Integration (III) with existing equipment, especially in brownfield sites. (not greenfiled); Brownfield integration: Machines: including OPC-UA, DDS, and MODBUS, as well as TCP-based sockets. Data: including for time series, location, ERP, and CRM systems. Programming language: provided for Java, Node.js, Python, Artifactory, GitHub, JaCoCo, and Ruby on Rails (RoR). Analytics: provided for Java, Matlab, and Python. Mobile devices: support for HTML5, browsers, smartphones, and tablets. Additional services: lifestyle, demographic insight, 911 precesses, incorporate local tax tae, payment and payroll application, etc. Developing Smarter to Innovate Faster: needs unique requirements that separate from traditional IT apps. typically 80% of time integrating and upgrading systems, vs. only 20% of time spent on innovation. Predix microservices: reusable software modules as building blocks to rapidly crate applications. enabling small teams of developers to version existing Connectivity, Asset, Field Agent, and Time Series, incrementally. allowing frequent releases for use while keeping the rest of the system available. >Top Continuous development using DevOps (Development & Operations) tools: provide tight integration including development, quality assurance, and IT. shorten development cycles and make agile and frequent user feedback. enables continuous development, where a new module can be automatically rolled into production in faster at a lower cost. Predix offers a rich development environment; placing user-centric: helps users visualize data in a way that is contextually relevant. helps answer the nagging questions related to application commercialization such as: What is an application's adoption rate? How are users using it? What features are not being use and why? What is the best subscription strategy? considers lifetime value of each individual subscriber. Cost dynamics: to meter the service so that the cost-to-service model is transparent.

2. エッジ〜クラウド・プラットフォーム Edge - Cloud方式: Predix Edgeの役割: Predix Machine, Predix接続, Predix EdgeManager ソフト定義マシン Predixマシンはセンサ･コントローラ･ゲートウェイ･ローカル機器で稼働 端末機器のセキュリティ･認証･ガバナンスを管理 稼働中のエッジ分析 End user -Egress 出口 -EdgeA (Cashe) -Intermediate trafic -EdgeB (PoP, Point of Presence) -Ingress 入口 -Origin Predixの接続性: 接続設定には6-12ヶ月 端末･ゲートウェイ･クラウド間通信 プロトコル非依存ネット接続 中央管理によるQoS･帯域の最適化 クラウドとローカル間データ伝達 携帯･固定･衛星通信での接続 セキュアVPN接続 端末管理にVNC/RDP/SSH/HTTP提供 端末間モニタリング ワンストップ報告 ZTP デフォルト設置のまま Predix EdgeManagerの機能: 端末機器管理の容易化 端末機器の状況･接続確認 端末機器の接続･断絶措置 Predixの特徴: Cloud Foundry (OSSベースPaaS) グローバルなアクセスポイント (Tier ⅡまたはTierⅢ) ソフト定義インフラ きめ細かい設備準備管理 高度なセキュリティ管理 (暗号化･鍵管理･トラブル事象対応･ログ･ネットワーク管理･端末間連携管理･24時間運用) 資産のモデル化: データ収集･処理･管理 迅速なアクセス･分析 セキュアな多ユーザモデル GE/非GE製センサ接続 リアルタイムでのデータ収集 パイプラインプロセス: 大量データ収集とフォーマット変換 データ管理 (大容量データ･RDBSM等の蓄積) 監察から対策実施へ 運用分析: 端末ではリアルタイムでのデータ分析 過去データ(ペタバイト規模) の分析 記述分析: 過去発生した事象とその原因の記述 予測分析: モデルに基づく予測事象 規範的分析: 意思決定の改善につながる分析 関連するユーザ体験(UX)の構築 context-adaptive≒context sensitive, 現在操作の場面に応じて >Top <TimescaleDB> TimescaleDB (TSDB): discrete-time data, like tides, sunspots, or stock prices; are called profiles, curves, or traces. Many RDB are often not modelled correctly to time series data. OSS time-series DB under Apache2.0 license. engineered up from PostgreSQL. introducing horizontal scale-out, automatic time/space partitioning. >Top PLC: Brownfiled Integration: PLC (Programmable Logic Controller): OPC Unified Architecture (OPC=OLE for Process Control, OLE=Object Linking & Embedding); Industrial interoperability to fulfil 1) utilization, 2) connection, 3) communication, and 4) security. DDS (Digital Data Storage); MODBUS: serial communication protocol developed by Modicon. RoR (Ruby on Rails); based on MVC (Model View Controller) MVC (Model View Controller): developed with Smalltalk by PARC in 1979, mostly used for GUI DevOpsによる顧客視点での連続開発環境: 商用化に当たっては、 以下等の回答支援 アプリ採用比率 ユーザの利用程度 利用されない特徴の抽出とその理由 ユーザ購読戦略 ユーザにとってのライフタイム価値 価格政策: 透明な費用対サービスモデル

>Top 3. Industrial-grade Security: >Top Predix adopted ISO 27001/27002 based IMS (Information Security Management and the Cloud Security Alliance-based Common Controls Matrix (CSA-CCM) including following: CSA/CCM 3.01: The Cloud Security Alliance Cloud Controls Matrix (CCM) provide fundamental security principles of a cloud provider. CCM also provides a customized relationship with ISO 27001/002, ISAC COBIT, PCI, NIST, PCI, and NERC CIP. >Top ISO 27001/002: requirements for establishing, implementing, maintaining, and improving information security within the organization. SOC2 Type1: developed by The American Institute of Certified Public Accountants (AICPA), a Service Organization Controls (SOC) report provides insight on internal controls and risks regarding services provided by a third-party service organization; generate a point-in-time assessment reporting on the fairness of management's description of the processes and design of the controls. SOC2 Tppe2: reports on fairness of management's description of the processes and design of the controls, throughout a specified period. HIPPA (protects): The Health Insurance Portability and Accountability Act (HIPAA) protects the privacy of individually identifiable health information. Export Controls/ITAR (International Traffic in Arms Regulations): US government regulates the transfer of information, commodities, technology, and software considered to be strategically important to US. Non-compliance with export controls can result in penalties.

3. 産業グレード･セキュリティ: IMS and the Cloud Secuirity: CCM: Cloudプロバイダの規制 ISO 27001/002: 組織内の情報セキュリティ SOC Type1: 第三者提供サービスの一時的な内部統制 SOC Type2: 同上の特定期間での内部統制 HIPAA: 個人特定可能な医療情報保護 ITAR: 米国の安全に係わる武器等の情報輸出規制

>Top 4. Digital Twin: DevOpsSec (Development Operations-Security) process for all apps and microservices; make tools available to help developers create secure workflows, handle data properly, evaluate app users, and dynamically test applications and APIs. to detect any abnormal behavior; reducing the possiblility of malware making its way into the run-time environment. Continuous monitoring: monitors at every layer, with data loss protection and malwre detection from external networks. creates a 'heat-map' dashboard form the Predix Security Operations team to protect customers. >Top Digital Twin: is a container for all knowledge and expertise about a specific asset class; to build apps about performace, optimization, and business transformation. provides knowledge, insights, and outcomes about its physical twin performance and operation (past and present), and allows for predictions ans simulation of future operation. AI or machine learning can also be added to the picture.

4. デジタル･ツイン: ヒートマップ: Digital Twin: コンピュータ上でリアルの動きを再現 製造ラインを変更する場合、コンピュータ上でも実際のしむレーションを実施することで開発期間・コスト削減 Cocurrent Engineeringを実現 フィジカルの世界をデジタルの世界にコピーして現状把握、故障予知、改善、次期モデル設計が可能

>Top 5. Growing the Ecosystem: Predix; the vision is bigger than just one company. Through partnerships with other technology companies, academia, consultants, and systems integrators, GE is sharing its expertise and know-how and co-innovation to drive important advances in functionality. >Top GE is a founding member of the Industrial Internet Consortium (IIC), an open membership NPO to share best practices, reference architecture, and case studies; and influencing global standards development to ensure interoperability. Oil & Gas; Mining; Power Generation; Power Distribution Manufacturing; Aviation; Transportation Intelligent Cities; Lighting; Water; Healthcare

5. エコシステムの成長: ハイテク会社、学界、コンサルタント、SIとの連携によるエコシステム構築 全米産業インターネット･コンソーシアム(IIC)の中核メンバーとして支援

>Top 6. xxxx:

6. xxxx:

Comment

Pretex of GE is the key to understand IoT or the Industrial Revolution 4.0. We needs to understand the Gemini of IoT age; the twins are composed of mortal Castor the son of human beings and immortal Pollux the son of computer.

GEのPretexはIoTあるいは第４次産業革命を理解する上でのキーとなる。 我々はIoT時代の双子座を理解することが必要である。人間の息子である命に限りのあるカストールとコンピュータの息子である不死のポルックスの双子である。