3. Microsoft Azure = オープンクラウド
3
+Hundreds of community
supported images on
VM Depot
SQL Server
オープンかつ柔軟なクラウドで幅広い選択肢をご提供してきました
Web App Gallery
Dozens of .NET & PHP
CMS and Web apps
Microsoft Azure
Azure 内の3分の1の
仮想マシンが
Linuxで稼働
4. Platform Services
Infrastructure Services
Web Apps
Mobile
Apps
API
Management
API Apps
Logic Apps
Notification
Hubs
Content
Delivery
Network (CDN)
Media
Services
BizTalk
Services
Hybrid
Connections
Service Bus
Storage
Queues
Hybrid
Operations
Backup
StorSimple
Azure Site
Recovery
Import/Export
SQL
Database
DocumentDB
Redis
Cache
Azure
Search
Storage
Tables
Data
Warehouse Azure AD
Health Monitoring
AD Privileged
Identity
Management
Operational
Analytics
Cloud
Services
Batch
RemoteApp
Service
Fabric
Visual Studio
App
Insights
Azure
SDK
VS Online
Domain Services
HDInsight Machine
Learning
Stream
Analytics
Data
Factory
Event
Hubs
Mobile
Engagement
Data
Lake
IoT Hub
Data
Catalog
Security &
Management
Azure Active
Directory
Multi-Factor
Authentication
Automation
Portal
Key Vault
Store/
Marketplace
VM Image Gallery
& VM Depot
Azure AD
B2C
Scheduler
The Azure Platform
10. 気になるお値段…
10
Amazon Azure Azure Google IDCF Sakura
Gen Kepler Kepler Pascal Kepler Pascal Pascal
GPU K80 x 1 K80 x 1 P100 x 1 K80 x 1 P100 x 1 P100 x 1
CPU Core 4 6 6 16 56 8
RAM 61GB 56GB 112GB 60GB 256GB 128GB
Cost / Hour $0.9 $0.9 まだ極秘 $0.7 $3.94 $3.19
https://aka.ms/gpupreview
11.
12. メジャーな分析ツールが
インストール、構成済み
Azure GPU VM 上に
Deep Learning 拡張機能が組み込み
Developer Editions of SQL & R Server
Azure Batch が利用可能に
Azure Data Science Virtual Machine
• 深層学習関連Azure サービス紹介 3分 o Azure全体像ちょっとだけ。 o 既存K80 o P40 / P100くるよ in this summer!Infiniband あるよ o Azure Data Science VMでサクッと始められる o Azure AI Batch Service!Azureへのバーストや、Hyper ParameterのGrid Search が可能になるよ• Azure ♡ Chainer 5分 o Chainer はDSVMですぐに始められる、今後、DSVMにPre install 予定 o Chainer MN 使うには、基本Azure オンプレでInfiniBandセットアップできますか?かなりつらい。 現在鋭意パフォーマンスチューニング中…(CentOS環境が今週出るそうなので、それが公開できるか。) ARM Templateのご紹介(間に合うか)• Microsoft ♡ PFN 2分 o 簡単に全体像の説明 o トレーニングプログラム、コミュニティの紹介
直近では、新規VMの3台に1台がLinuxで作成されています。(2016 TR23での発表)
These three horsemen bring incredible power.
1. Your data with any AI tools such as Chainer
2. Training ith Scale-Out GPU Clusters on Demand
3. Power of ADL with Cognitive Service APIs
The basic premise of IntelligenceDB pattern is that you “push intelligence to where your data lives”. When you do this with an industrial strength database engine like SQL Server, you can get throughput, parallelism, security, reliability, compliance certifications and manageability, all in one. It’s a big win for developers – you don’t have to build it separately.
Furthermore, just like data in databases can be shared across multiple applications, you can now share the predictive models. Models and intelligence beome “yet another type of data”, managed by the DBMS.
1. Bring intelligence to where your data lives
2. On the most trusted and performant plat
3. With any language, any platform, anywhere
A few weeks ago, we announced SQL Server 2017 CTP 2.0, the first commercial database with AI built-in.
By adding Python support in addition to R and adding real-time scoring capabilities, now you can run machine learning models directly in SQL Server to eliminate the need to move data, increase efficiency and help uncover new insights.
You can easily incorporate AI models into SQL queries, allowing you to infuse your applications with intelligence with minimal extra coding.
It also supports graph data, enabling efficient analysis of complex relationships.
And the database server uses machine learning internally to adaptively process queries for best possible performance.
It Supports Linux distributions including RedHat Enterprise Linux (RHEL), Ubuntu, and SUSE Enterprise Linux (SLES)
You can run SQL Server in Windows and Linux containers on Docker
It’s an amazing harness for AI applications.