Worldwide Scalable and Resilient Messaging Services by CQRS and Event Sourcin...DataWorks Summit
ChatWork is one of major business communication platforms in Japan. We keep growing up for 5+ years since our service inception. Now, we hold 110k+ of customer organizations which includes large organizations like telecom companies and the service is widely used across 200+ countries and regions.
Nowadays we have faced drastic increase of message traffic. But, unfortunately, our conventional backend was based on traditional LAMP architecture. Transforming traditional backend into highly available, scalable and resilient backend was imperative.
To achieve this, we have applied “Command Query Responsibility Segregation (CQRS) and Event Sourcing” as a heart of its architecture. The simple idea of segregation brings us independent command-side and query-side system components and it can subsequently achieve highly available, scalable and resilient systems. It is desirable property for messaging services because, for example, even if command-side was down, user can keep reading messages unless query-side was down. Event Sourcing is another key technique to enable us to build optimized systems to handle heterogeneous write/read load. This means that we can choose optimized storage platform for each side. Moreover, the event data can be the rich source for real-time analysis of user’s communication behavior. We have chosen Kafka as a command-side event storage, HBase as a query-side storage, Kafka Streams as a core library to give eventual consistency between the two sides. In application layer, Akka has been chosen as a core framework. Akka can be a good choice as an abstraction layer to build highly concurrent, distributed, resilient and message-driven application effectively. Backpressure introduced by Akka Stream can be important technology to prevent from overflow of data flows in our backend, which contributes system stability very well.
In this session, we talk about how above architecture works, how we concluded above architectural decisions on many trade-offs, what was achieved by this architecture, what was the pain points (e.g. how to guarantee eventual consistency, how to migrate systems in the real project, etc.) and several TIPS we learned for realizing our highly distributed and resilient messaging systems.
ChatWork is a business communication platform for global teams. Our four main features are enterprise-grade group chat, file sharing, task management and video chat. NTT DATA is one of biggest solution provider in Japan and providing technical support about Open Source Software and distributed computing. The project has been conducted with cooperation of ChatWork and NTT DATA.
NTTコミュニケーションズでは、Azure Stack Hub with GPUを先行で導入し検証を行っています。本資料では、実際に利用している立場からデモを交えつつAzure Stack Hub with GPUのユースケースをお話すると共に、GPUのベンチマークを含む他社クラウドとの性能比較結果について情報共有をいたします。
Worldwide Scalable and Resilient Messaging Services by CQRS and Event Sourcin...DataWorks Summit
ChatWork is one of major business communication platforms in Japan. We keep growing up for 5+ years since our service inception. Now, we hold 110k+ of customer organizations which includes large organizations like telecom companies and the service is widely used across 200+ countries and regions.
Nowadays we have faced drastic increase of message traffic. But, unfortunately, our conventional backend was based on traditional LAMP architecture. Transforming traditional backend into highly available, scalable and resilient backend was imperative.
To achieve this, we have applied “Command Query Responsibility Segregation (CQRS) and Event Sourcing” as a heart of its architecture. The simple idea of segregation brings us independent command-side and query-side system components and it can subsequently achieve highly available, scalable and resilient systems. It is desirable property for messaging services because, for example, even if command-side was down, user can keep reading messages unless query-side was down. Event Sourcing is another key technique to enable us to build optimized systems to handle heterogeneous write/read load. This means that we can choose optimized storage platform for each side. Moreover, the event data can be the rich source for real-time analysis of user’s communication behavior. We have chosen Kafka as a command-side event storage, HBase as a query-side storage, Kafka Streams as a core library to give eventual consistency between the two sides. In application layer, Akka has been chosen as a core framework. Akka can be a good choice as an abstraction layer to build highly concurrent, distributed, resilient and message-driven application effectively. Backpressure introduced by Akka Stream can be important technology to prevent from overflow of data flows in our backend, which contributes system stability very well.
In this session, we talk about how above architecture works, how we concluded above architectural decisions on many trade-offs, what was achieved by this architecture, what was the pain points (e.g. how to guarantee eventual consistency, how to migrate systems in the real project, etc.) and several TIPS we learned for realizing our highly distributed and resilient messaging systems.
ChatWork is a business communication platform for global teams. Our four main features are enterprise-grade group chat, file sharing, task management and video chat. NTT DATA is one of biggest solution provider in Japan and providing technical support about Open Source Software and distributed computing. The project has been conducted with cooperation of ChatWork and NTT DATA.
NTTコミュニケーションズでは、Azure Stack Hub with GPUを先行で導入し検証を行っています。本資料では、実際に利用している立場からデモを交えつつAzure Stack Hub with GPUのユースケースをお話すると共に、GPUのベンチマークを含む他社クラウドとの性能比較結果について情報共有をいたします。
CloudNative Days Tokyo 2021で発表した資料です。
https://event.cloudnativedays.jp/cndt2021/talks/1279
Terraform、Pulumi、Kustomize、CrossplaneなどといったInfrastructure as Codeを取り巻くエコシステムを分析し、パブリッククラウドやKubernetesの力を最大限に引き出すためのツールスタックをどう組み上げていくか考察しています。
Slides at OpenStack Summit 2017 Sydney
Session Info and Video: https://www.openstack.org/videos/sydney-2017/100gbps-openstack-for-providing-high-performance-nfv
Slide at OpenStack Summit 2018 Vancouver
Session Info and Video: https://www.openstack.org/videos/vancouver-2018/can-we-boost-more-hpc-performance-integrate-ibm-power-servers-with-gpus-to-openstack-environment
This slide was for CLOUDEXPO 2017 in NYC. Consists of two part, One is for introducing existing WebRTC - IoT use cases. Another is conceptual consideration of Edge Computing scenario which leveraging WebRTC technology.
RabbitMQ is said a point of bottleneck in OpenStack.
We researched RabbitMQ and analyzed OpenStack RPC messaging.
This slide shows that RabbitMQ can scale out with HA setting.
More from NTT Communications Technology Development (20)
【DLゼミ】XFeat: Accelerated Features for Lightweight Image Matchingharmonylab
公開URL:https://arxiv.org/pdf/2404.19174
出典:Guilherme Potje, Felipe Cadar, Andre Araujo, Renato Martins, Erickson R. ascimento: XFeat: Accelerated Features for Lightweight Image Matching, Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (2023)
概要:リソース効率に優れた特徴点マッチングのための軽量なアーキテクチャ「XFeat(Accelerated Features)」を提案します。手法は、局所的な特徴点の検出、抽出、マッチングのための畳み込みニューラルネットワークの基本的な設計を再検討します。特に、リソースが限られたデバイス向けに迅速かつ堅牢なアルゴリズムが必要とされるため、解像度を可能な限り高く保ちながら、ネットワークのチャネル数を制限します。さらに、スパース下でのマッチングを選択できる設計となっており、ナビゲーションやARなどのアプリケーションに適しています。XFeatは、高速かつ同等以上の精度を実現し、一般的なラップトップのCPU上でリアルタイムで動作します。
セル生産方式におけるロボットの活用には様々な問題があるが,その一つとして 3 体以上の物体の組み立てが挙げられる.一般に,複数物体を同時に組み立てる際は,対象の部品をそれぞれロボットアームまたは治具でそれぞれ独立に保持することで組み立てを遂行すると考えられる.ただし,この方法ではロボットアームや治具を部品数と同じ数だけ必要とし,部品数が多いほどコスト面や設置スペースの関係で無駄が多くなる.この課題に対して音𣷓らは組み立て対象物に働く接触力等の解析により,治具等で固定されていない対象物が組み立て作業中に運動しにくい状態となる条件を求めた.すなわち,環境中の非把持対象物のロバスト性を考慮して,組み立て作業条件を検討している.本研究ではこの方策に基づいて,複数物体の組み立て作業を単腕マニピュレータで実行することを目的とする.このとき,対象物のロバスト性を考慮することで,仮組状態の複数物体を同時に扱う手法を提案する.作業対象としてパイプジョイントの組み立てを挙げ,簡易な道具を用いることで単腕マニピュレータで複数物体を同時に把持できることを示す.さらに,作業成功率の向上のために RGB-D カメラを用いた物体の位置検出に基づくロボット制御及び動作計画を実装する.
This paper discusses assembly operations using a single manipulator and a parallel gripper to simultaneously
grasp multiple objects and hold the group of temporarily assembled objects. Multiple robots and jigs generally operate
assembly tasks by constraining the target objects mechanically or geometrically to prevent them from moving. It is
necessary to analyze the physical interaction between the objects for such constraints to achieve the tasks with a single
gripper. In this paper, we focus on assembling pipe joints as an example and discuss constraining the motion of the
objects. Our demonstration shows that a simple tool can facilitate holding multiple objects with a single gripper.