Oracle Engineered Systems Christian Kühne ServerTechnologyCustomerCenter – FusionMiddleware
Safe Harbor Statement The following is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle.
© Copyright 2012, Oracle Corporation
Oracle – vor & nach SUN
Oracle Software – Software Sicht jedes OS, jede HW
Generische Software-Sicht Funktionale Betrachtung Software in Betrieb bringen Begrenzte Optimierung nach „unten“ Häufig ausreichend
Nicht funktionale Betrachtung Software in Betrieb halten Ausfall / Verfügbarkeit Performance / Skalierung Patchen / Management Abhängigkeiten, zB Libraries → OS Treiber → Platte
Problem- / Errorhandling Strom / Klima
Ziel-Systemarchitektur Das „well balanced system“
Storage I/O
Well Balanced System Processors
Interconnect Memory
Realitätscheck: Unausgewogene Komponenten, falsche Konfiguration, Engpässe … irgendwo – und sich ändernde Rahmenbedingungen! 6
Typische technische Architektur 7 Racks, 128 Verbindungen, bis zu 7 Teams und ‘Vendors’ Ethernet Interconnect
Network Network Team Team
Switch Switch Vendor Vendor
Database
DBAs DBAs
DB DB Vendor Vendor
Operating System
Unix Unix Admin Admin
OS OS Vendor Vendor
HBA
HW HW Admin Admin
HBA HBA Vendor Vendor
Volume Manager
Storage Storage Design Design
LVM LVM Vendor Vendor
FC Switches
Data Data Fabric Fabric
FC FC switch switch Vendor Vendor
LUNS
Storage Storage Admin Admin
Storage array
Vendor Vendor Support Support
Storage Storage Vendor Vendor
7
Challenges of Today’s IT Systems
Applications Middleware Database Operating System Virtual Machine Servers Storage
Oracle Engeineered Systems
Oracle Engineered Systems • optimiert und zwischen Hard-
und Software feinabgestimmt • standardisiert • ermöglicht schnelle Aufstellung und Einrichtung • reduziert Aufwand für Betrieb
Standardisiert Eine Standard-Platfform für alle Lastprofile • alle Maschinen sind identisch • vorkonfigurierte, getestete und durch Support unterstützte Konfiguration • identische Umgebung in der Oracle Entwicklungsabteilung
• zum Betrieb vorhandener Anwendungen - gesamtes Fusion Middleware Portfolio - jede Linux 5 oder Solaris 11 Anwendung - keine eigene Zertifizierung notwendig
• nutzt das Oracle-Ökosystem - Skills, Erfahrungen, Mitarbeiter, Partner
© 2010 Oracle Corporation
11
Vollständiges und integriertes Management Wartung Fernwartung Phone Home My Oracle Support Integration
Deployment Enterprise Manager Grid Control
Provisionierung von Firmware, OS, Middleware, und Anwendungen Cloning und Scale-out
OpsCenter
Management
Test
End-to-End Diagnose Konfigurations-Management Patch Automation
Funktionstests Lasttests Test Management
Überwachung Application-to-Disk Exalogic Monitoring Integration Stromverbrauch
© 2010 Oracle Corporation
12
Enterprise Manager Vom Applikations- bis zum Datacenter-Betrieb Enterprise Manager Grid Control
Enterprise Manager OpsCenter Im Kontext
ApplikationsBetrieb
DatacenterBetrieb
GC Agent
ILOM
Exalogic
Integriertes „Lights Out“ Management Betriebsfähigkeit Eingebaut
OpsCenter
•
Proaktive Betriebsfähigkeit “Phone Home”
OpsCenter – –
My Oracle Support
•
ILOM Integration Direkte Integration mit My Oracle Support
Exalogic ILOM – – – –
Exalogic
Compute Nodes Gateways Storage Switches
EnterpriseManager12c Cloud Control
ZFS Storage Appliance
Cloud Management Complete Cloud Lifecycle Management ● ● ● ●
Creating virtual system slices Self-service assembly deployment Automatic scale-up and scale-down Metering and chargeback
Deep, Integrated Hardware, Software and Applications Management Proactive Support ● ● ●
Phone home Patch recommendations Health checks
Skalierung von der einzelnen Anwendung bis zur Enterprise Cloud
Multi-rack Full Rack Half Rack Quarter Rack Eighth Rack
Exadata Database Machine
Exadata ist ‘Smart Storage’
INFINIBAND
Rechenintensive Verarbeitung
Bandbreitenintensives Suchen & Lesen
20
© Copyright 2012, Oracle Corporation
Technische Daten der Exadata Beispiel: Full Rack
Datenbank Server • 8 RAC Knoten, Linux/Solaris
Intelligent Storage
• 2 Intel Sockets 6 Cores
• 14 Storage Server, Linux
• 2 Xeon CPUs per server
InfiniBand Netzwerk
• 100TB High Performance
• Redundante 40Gb/s Switches • Einheitliches Netzwerk für Server & Storage
• 504 TB High Capacity • 12TB/h Datenladeperformance • 5 TB PCIe Flash
redundant ausgelegt! 22
Exadata Smart Flash Cache Extreme Performance
1‘500‘000 IO/s
5‘000 Disks 168 Disks
23
Exalogic Elastic Cloud
Exalogic “Design Center” Exalogic Konstruktionsprinzipien • Modernste Sun Oracle Hardware Komponenten • Dense Compute Cluster • High Performance Networking • Balanced System Design – CPU/Memory/Netzwerk/Storage • Generelle Anwendbarkeit für Mid-Tier Workload • Plattform für heterogene Konsolidierung
Exalogic High Level System Architektur InfiniBand Network (40 Gbs QDR)
Oracle Database (non-Exadata)
Datacenter Client Network (10GbE)
X4170M2 Compute Servers (4 .. 30)
NM2-GW 10GbE-IB Gateways
7320 Storage
Datacenter Management Network (1GbE)
Cisco Mgmt Switch
Exalogic Elastic Cloud
Exadata Database Machine
Exalogic Hardware 4 .. 30 X4170 M2 • 2 x 6 core CPU → 48 .. 360 Cores • 12 x 8 GB DIMM → 384 .. 2880 GB RAM • 2 x 32 GB SATA-SSD (RAID 1) 7320 ZFS Storage Appliance • 20 x 2 TB drives • 2 drives hotspare (typical) • → 36 TB disk (18 TB in RAID 1) Infiniband 40 GB/s 10 GB Ethernet for external connections 1 GB Ethernet for iInternal management network
Exalogic – “Cluster” Sicht
• • •
•
•
• •
•
Erweiterbare Standard “Racks” Eighth, Quarter, Half und Full Kombinierung vieler Racks zu einem Cluster, basierend auf eine gemeinsame InfiniBand Fabric Bis zu 8 Racks ohne zusätzliche, externe Netzwerk Hardware Cluster mit mehr als 8 Racks verwenden externe, zusätzliche (InfiniBand) Switches Exadata hat eine ähnliche Mulit-Rack Strategie Exalogic und Exadata Racks können kombiniert werden und verwenden die gleiche InfiniBand Fabric Aus der Fabric Sicht, stellen sie eine „Maschine“ dar
Optimiert für Enterprise Java, nutzbar für jede Software Oracle Anwendungen
3rd Party Anwendungen
Extreme Performance
Enterprise Java, Fusion Middleware 11g
Eigenentwicklung Zertifiziert
Oracle Linux oder Solaris Anwendungen
Oracle Exalogic Elastic Cloud
Basis: Oracle WebLogicServer • Geclusterte Server
für Hochverfügbarkeit und Skallierbarkeit • Administration „per domain“
A1
A1
A1
A1
A2
A2
A2
A2
Admin Server Managed Server Node Manager
© Copyright 2012, Oracle Corporation
Java Performance • Network I/O optimiert
für InfiniBand • Vermeidung von buffer copies • größere packet size zur Reduzierung von Overhead im Netzwerk • Automatische Optimierung des Thread-Pool (Anpassung an Cores)
Exalogic ExalogicSoftware Software CP/IP CP/IP TPoIB TPoIB
Sockets Sockets Direct Direct Protocol Protocol
InfiniBand InfiniBandCore Core Host HostChannel Channel Adapter Adapter(Driver) (Driver)
Inter-Process-Communication
Faster Network Fewer, Larger Messages Parallel Muxer
Parallel Muxer
WebLogic
SDP over InfiniBand
Exalogic
WebLogic
Lesson: Understanding the Sockets Direct Protocol • For high performance computing environments, the capacity to move data
across a network quickly and efficiently is a requirement. Such networks are typically described as requiring high throughput and low latency. High throughput refers to an environment that can deliver a large amount of processing capacity over a long period of time. Low latency refers to the minimal delay between processing input and providing output, such as you would expect in a real-time application. • In these environments, conventional networking using socket streams can
create bottlenecks when it comes to moving data. Introduced in 1999 by the InfiniBand Trade Association, InfiniBand (IB) was created to address the need for high performance computing. One of the most important features of IB is Remote Direct Memory Access (RDMA). RDMA enables moving data directly from the memory of one computer to another computer, bypassing the operating system of both computers and resulting in significant performance gains. • ...
http://download.oracle.com/javase/tutorial/sdp/sockets/index.html
Roadmap
Exabus - High Speed Network Virtualization Direct Memory I/O for Java and C++
Standard Hardware I/O Application
Application
Application Buffer
Copy
Exabus Application
Application
Application Buffer
TCP IP Transport Kernel
20% Buffer Copies 40% Transport Processing 40 % Kernel Context Switches
Zero Buffer Copy Direct Memory Access Kernel Bypass
4X Throughput, 6X Lower Latency
Oracle Exalogic Elastic Cloud World’s Best Foundation for Cloud
• optimiert für Java Performance • Anwendungen laufen bis zu 10x schneller • ausgerichtet auf unternehmens-
kritische Systeme • optimal für Anwendungskonsolidierung • ausgezeichnete Anwendungsskalierbarkeit
• “engineered System” • geringe Kosten für Einrichtung und Betrieb • schnelle Implementierung
Beispiele & UseCases
UseCase – Exa* Exadata
Exalogic
• Enterprise database
• Extreme perfomance
consolidation • Data Warehousing • OLTP
• Applications • Fusion Midleware, SOA,
BPM • Java Applications • consolidation
Exadata Exadata&&Exalogic Exalogic
UseCase – große Datenmengen • Schnelle Zugriff auf große
Datenmengen • Relationale Daten in Exadata • Bearbeitung mit Java in Exalogic • Beide Maschienen innerhalb derselben Infiniband-Fabric
Exalogic Elastic Cloud
Exadata Database Machine
NIST Definition von Cloud Computing Cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. This cloud model promotes availability and is composed of: 5 Essential Characteristics • On-demand self-service • Resource pooling • Rapid elasticity • Measured service • Broad network access Source: NIST Definition of Cloud Computing v15
3 Service Models • SaaS • PaaS • IaaS
4 Deployment Models • Public Cloud • Private Cloud • Community Cloud • Hybrid Cloud
ExaCloud: Own The Private Cloud
Fusion Applications
3rd Party Applikationen
• •
Transformation in IT-As-A-Service Vorgefertigte “Private Cloud in a Box”, geringe TCO
UseCase – Cloud Plattform Beispiel: PaaS Third Third Party Party Applications Applications
Oracle Oracle Applications Applications
ISV ISV Applications Applications
Platform as a Service Cloud Management
Shared Services
Oracle Enterprise Manager Integration
Process Mgmt
Security
User Interaction:
WebCenter
Application Grid Database Grid
Oracle Operating SolarisSystems: Oracle Enterprise Oracle Enterprise Linux Linux Oracle VM for x86 Servers Storage
Lifecycle Management Application Performance Management
Infrastructure as a Service Oracle VM for SPARC (LDom) Solaris Containers
Configuration Mgmt
Application Quality Management Ops Center Physical and Virtual Systems Management
A New Model for Cloud Application Deployment Autonomous, Dynamic, Simple
Deployment Plan and Software
Copyright © 2011 Oracle Corporation
Virtual Assembly
Elastic Cloud
Oracle Virtual Assembly Builder Automating deployment of ‘assemblies’ config 1
config 2
Assembly = Dev/Test Environment
Appliances (VM Templates + configuration Metadata) + relationships & start order Metadata
Production Environments
Package up complex structure from dev/test and reconstitute in production Minimize setup time and risk of hard-to-debug configuration errors
• •
Easily replicate in production with minor variations Each instance has well-contained configuration parameters for flexibility
• •
UseCase – complex event procesing • Schnelle Bearbeitung komplexer Events • Umfangreiche (komplexe) Regeln in RDBMS
© Copyright 2012, Oracle Corporation
Exalogic: World’s Best Foundation for Cloud Hardware and Software Engineered to Work Together
Building the 21st Century Datacenter
Q& 42 © Copyright 2012, Oracle Corporation