Windows Azure Training Tokens You should already have

  • Slides: 110
Download presentation
Windows Azure ‘Training’ Tokens • You should already have received an email with the

Windows Azure ‘Training’ Tokens • You should already have received an email with the login information • Good for 1 month Expires May 8 th • No credit card; no commitments! • Meant for ‘trying out’ Windows Azure • After 1 month, transition strategy handled by regional Microsoft representative

Azure Access Requirements • Compute Hours • Data Storage – Blob – SQL Azure

Azure Access Requirements • Compute Hours • Data Storage – Blob – SQL Azure – ‘Dallas’ data • As I talked to faculty last night – Don’t know – Want accounts for 4 students to probe further [Research] – Accounts for my class of 20 students [teaching] • Solution offering – a) Highest denominator? b) Mean? c) Least common denominator? • So think of ’ 30 -day’ account [ends may 8 th] as a way to learn the Azure features and what aspects your work needs. • Let your Microsoft contact know what your needs are in 1. Research 2. Teaching

Architecting for the Windows Azure Platform Krishna Kumar krishna. kumar@microsoft. com http: //azurepilot. com

Architecting for the Windows Azure Platform Krishna Kumar krishna. kumar@microsoft. com http: //azurepilot. com Twitter: Krishna. On. Azure

http: //Azure. Pilot. com

http: //Azure. Pilot. com

State of Cloud Computing • Perceptions – “The end of software” – On-demand infrastructure

State of Cloud Computing • Perceptions – “The end of software” – On-demand infrastructure – Cheaper and better • Reality • > Types • • • Public Private Internal External – Hybrid world; not “all-or-nothing” Hybrid – Leverage existing IT skills and > Categories investments • Saa. S – Seamless user experiences • Paa. S – Evolutionary; not revolutionary • Iaa. S Drivers – Ease-of-use, convenience – Product effectiveness – Simplify IT, reduce costs

IT as a Service Infrastructure Applications Runtimes (as a Service) Applications Runtimes Security &

IT as a Service Infrastructure Applications Runtimes (as a Service) Applications Runtimes Security & Integration Databases Servers Virtualization Server HW Storage Networking Databases Virtualization Server HW Storage Networking Managed by vendor Security & Integration Managed by vendor You manage Security & Integration (as a Service) You manage (On-Premise) Platform You manage Private

Microsoft in the Cloud (15 years) 450 M+ active users (13 years) 550 M

Microsoft in the Cloud (15 years) 450 M+ active users (13 years) 550 M users/mth (11 years) 320 M+ (12 years) active Largest nonusers ICP/IP cloud service x 100 M users (11 years) 2 B queries/mth (15 years) 450 M+ active users (7 years) 5 B conf min/yr (6 years) 4 B emails/day

The Microsoft Cloud Categories of Services Application Services Software Services Platform Services Infrastructure Services

The Microsoft Cloud Categories of Services Application Services Software Services Platform Services Infrastructure Services

Workload Patterns Optimal For Cloud “Growing Fast“ Compute “On and Off “ Inactivity Period

Workload Patterns Optimal For Cloud “Growing Fast“ Compute “On and Off “ Inactivity Period Average Usage Average Time On & off workloads (e. g. batch job) Over provisioned capacity is wasted Time to market can be cumbersome Successful services needs to grow/scale Keeping up w/ growth is big IT challenge Complex lead time for deployment “Predictable Bursting“ Compute “Unpredictable Bursting“ Average Usage Time Unexpected/unplanned peak in demand Sudden spike impacts performance Can’t over provision for extreme cases Services with micro seasonality trends Peaks due to periodic increased demand IT complexity and wasted capacity

But what about these new scenarios?

But what about these new scenarios?

Consider these statistics: • EIGHTEEN months ago, Li & Fung 100 GB of information

Consider these statistics: • EIGHTEEN months ago, Li & Fung 100 GB of information flow through its network each day. Now the amount has increased tenfold • During 2009, American drone aircraft flying over Iraq and Afghanistan sent back around 24 years’ worth of video footage • When the Sloan Digital Sky Survey started work in 2000, its telescope in New Mexico collected more data in its first few weeks than had been amassed in the entire history of astronomy. Now, a decade later, its archive contains 140 TB of information. • A successor, the Large Synoptic Survey Telescope, due to come on stream in Chile in 2016, will acquire that quantity of data every five days. • Wal-Mart, a retail giant, handles more than 1 m customer transactions every hour, feeding databases estimated at more than 2. 5 PB — the equivalent of 167 times the books in America’s Library of Congress • Mankind created 150 exabytes (billion gigabytes) of data in 2005. This year, it will create 1, 200 exabytes.

But, How does it work?

But, How does it work?

A 2 -D Perspective Infrastructure App Dev • Hardware • • • – Servers

A 2 -D Perspective Infrastructure App Dev • Hardware • • • – Servers – Storage – Networks • Management Concurrency Coordination Fault Tolerance Open API formats Open Data Formats Data Intensive Computing

Windows Azure An illustration Application Compute Config . NET Services SQL Azure Applications Windows

Windows Azure An illustration Application Compute Config . NET Services SQL Azure Applications Windows Server Windows Vista/XP Windows Mobile Others Storage Fabric

Windows Azure is a place to run your applications.

Windows Azure is a place to run your applications.

Windows Azure is a cloud computing platform, so it’s designed to scale.

Windows Azure is a cloud computing platform, so it’s designed to scale.

Windows Azure is a utility computing platform, so you pay for what you use.

Windows Azure is a utility computing platform, so you pay for what you use.

Your Application

Your Application

Windows Azure does scale out.

Windows Azure does scale out.

Stateless compute + Durable storage --------------= Scalable application

Stateless compute + Durable storage --------------= Scalable application

Storage Service

Storage Service

Durable Storage Blobs Tables Queues … • Three replicas of everything • REST API

Durable Storage Blobs Tables Queues … • Three replicas of everything • REST API

Storage • Simple, essential storage abstractions: – Large items of data: Blobs, file streams,

Storage • Simple, essential storage abstractions: – Large items of data: Blobs, file streams, … – Service state: Simple tables, caches, … – Service communication: Queues, locks, … • With an emphasis on: – Massive scale, availability and durability – Geo-location and geo-replication • This is not a relational database in the cloud

Management tasks are automated by the Fabric Controller.

Management tasks are automated by the Fabric Controller.

You tell the Fabric Controller what to do, and it figures out how to

You tell the Fabric Controller what to do, and it figures out how to do it.

For this to work, you need to tell us about your application.

For this to work, you need to tell us about your application.

Defining Your Service • What are the components (roles)? • How do they interact?

Defining Your Service • What are the components (roles)? • How do they interact? • At what scale? LB Webrole Web role Workerrole Worker role

Putting It All Together LB Web role Worker role Storage

Putting It All Together LB Web role Worker role Storage

Developer Experience is Key • Development – Offline cloud simulation – Visual Studio integration

Developer Experience is Key • Development – Offline cloud simulation – Visual Studio integration • Maintenance – Local debugging – APIs for logging • Management – APIs for configuration management – Web portal

Geo-location • Windows Azure is in multiple “locales” • You can: – Choose a

Geo-location • Windows Azure is in multiple “locales” • You can: – Choose a locale for any of your applications – Create an “affinity group” to co-locate a set of applications from your cloud project • Available through the developer portal

Windows Azure Your Service D N S L B Web Portal (API) Fabric Controller

Windows Azure Your Service D N S L B Web Portal (API) Fabric Controller L B

Service Deployment Your Model Service D N S L B Web Portal (API) config

Service Deployment Your Model Service D N S L B Web Portal (API) config DNS Fabric Controller L B

Service Scaling Your Service D N S Service L B Service Web Portal (API)

Service Scaling Your Service D N S Service L B Service Web Portal (API) Model Fabric Controller L B

Service Monitoring & Recovery Your Service D N S Service L B Web Portal

Service Monitoring & Recovery Your Service D N S Service L B Web Portal (API) Model Service ! Fabric Controller Service L B

Putting It All Together LB n Web Role Cloud Storage m Worker Role

Putting It All Together LB n Web Role Cloud Storage m Worker Role

SQL Azure Formerly known as SQL Services SQL Azure Database Others (Future) . NET

SQL Azure Formerly known as SQL Services SQL Azure Database Others (Future) . NET Services SQL Azure Applications Windows Server Windows Vista/XP Windows Mobile Others

SQL Azure • Today: – SQL Azure Database • Formerly known as SQL Data

SQL Azure • Today: – SQL Azure Database • Formerly known as SQL Data Services (SDS) – And before that as SQL Server Data Services (SSDS) – “Huron” Data Hub • Built on the Microsoft Sync Framework • In the future: – Reporting – Analysis – Extract/Transform/Load (ETL) services – More

SQL Azure Database An illustration TDS Database SQL Azure Database “Huron” Data Hub Others

SQL Azure Database An illustration TDS Database SQL Azure Database “Huron” Data Hub Others (Future)

SQL Azure Database Using one or multiple databases SQL Azure Database Application Database

SQL Azure Database Using one or multiple databases SQL Azure Database Application Database

? . NET Services SQL Azure Applications Windows Server Windows Vista/XP Windows Mobile Others

? . NET Services SQL Azure Applications Windows Server Windows Vista/XP Windows Mobile Others ss Control Service Bus

Service Bus • The problem: Exposing internal applications on the Internet isn’t easy –

Service Bus • The problem: Exposing internal applications on the Internet isn’t easy – Network address translation (NAT) and firewalls get in the way • The solution: – Service Bus provides a cloud-based intermediary between clients and internal applications • Direct or queued – It also provides a service registry that clients can use to find the services they need

su. B ecivre. S gnitartsull. I ecivre. S rtsige. R niopdn. E revocsi. D

su. B ecivre. S gnitartsull. I ecivre. S rtsige. R niopdn. E revocsi. D )2 stniopdne ssecc. A )3 noitacilppa noitacilpp. A X noitazinagr. O lortno. C ssecc. A su. B ecivre. S

Application Models Web Hosting High Performance Computing § Massive scale infrastructure § Parallel &

Application Models Web Hosting High Performance Computing § Massive scale infrastructure § Parallel & distributed processing § Burst & overflow capacity § Massive modeling & simulation § Temporary, ad-hoc sites § Advanced analytics Application Hosting Information Sharing § Hybrid applications § Reference data § Composite applications § Common data repositories § Automated agents / jobs § Knowledge discovery & mgmt Media Hosting & Processing Collaborative Processes § CGI rendering § Multi-enterprise integration § Content transcoding § B 2 B & e-commerce § Media streaming § Supply chain management Distributed Storage § External backup and storage § Health & life sciences § Domain-specific services

Case Study Background § Web-based ERP and accounting software to help mid-sized businesses automate

Case Study Background § Web-based ERP and accounting software to help mid-sized businesses automate customized business processes § Business needs: • Launch a Saa. S offering which includes on demand infrastructure and billing • Maintain a 24/7 production environment • Minimize development efforts for both on-premise and Saa. S offerings § Technical implementation • Web – IIS 7, ASP. NET 3. 5 Solution § Cloud-based Saa. S deployment § Windows Azure § SQL Azure Benefits § Application migration completed in 3 weeks with 2 developers § Customers can implement solution without investing in onpremise infrastructure; can save ~$8, 000 for startup costs § Lowered capital expenditures • Server – SQL Server 2005/2008, Windows Server 2003/2008 § Improved time-to-market and agility in meeting customer needs • Tools – Visual Studio 2008, . NET Framework 3. 5 § More strategic use of IT resources

Case Study Background Solution § Online service providing productivity tools to help students conduct

Case Study Background Solution § Online service providing productivity tools to help students conduct research, write better papers, and get help on-demand § Hybrid cloud application § Technical implementation § Windows Azure application • Client – Windows Mobile 6 • Web – IIS 6, ASP. NET 3. 5 • Server – Windows Server 2003, SQL Server 2005 § Cloud-based extensions and services mash-ups with onpremise implementation §. NET Services service bus, access control, & workflows Benefits • Services – Bing Search, Windows Live Messenger Web Toolkit § $500, 000 savings in dev. costs • Tools – Visual Studio 2008, . NET Framework 3. 5, . NET Compact Framework 3. 5 § More effective development § 40% faster time-to-market § Cost-effective scalability and enhanced availability

Case Study Browser Internet Clipping Toolbar Full Featured Javascript Editor Local / Network File

Case Study Browser Internet Clipping Toolbar Full Featured Javascript Editor Local / Network File Folder JSON / AJAX/ HTTP IE 6/7/8* and Firefox • Rich Web 2. 0 Technologies: Javascript Java Ajax CSS XHTML F W — L B — C a c h e Eduify Workflows ASPX Pages Facebook Connect Integration Local Storage Eduify Business Objects Common Services Notification Logging Security SQL Server ASP. NET / CLR IIS Bing Search Integration for Citation Check Third Party Mashups

Case Study Live Presence IM Integration Connected w/ Services Search Citation / Plagiarism Mash-up

Case Study Live Presence IM Integration Connected w/ Services Search Citation / Plagiarism Mash-up with Microsoft Services Cloud-Based Extensions On-Premise Implementation ASP. NET Certain ASPX | Workflows | Service Bus | Access Controls Workflows Business Objects Common Services SQL Server Service Bus Local Storage

Case Study About Solution § kbb. com, established 1995; vehicle data aggregation and publication

Case Study About Solution § kbb. com, established 1995; vehicle data aggregation and publication service provider § 14 M UU/month § Multiple physical data centers § Cloud-based overflow capacity § Windows Azure Web Role § SQL Azure database Benefits § ~$100, 000 savings / year in hosting costs alone § Retire failover data center (27 Web servers & 9 SQL Servers) § <1% code changes needed for application compatibility § 6 weeks down to 6 minutes to add server capacity

Case Study About Solution § Legacy call-center application for order processing and fulfillment §

Case Study About Solution § Legacy call-center application for order processing and fulfillment § Cloud-based Web-enablement of on-premise assets § On-premise systems integration with ERP and master data management § Windows Azure Web Role § Web services integration via onpremise ESB endpoints Benefits § “In a matter of weeks, we were able to see the fruits of our efforts very rapidly. “ – Kevin Flowers, Dir. Enabling Technologies § “The resilience and reliability of this platform is some of the benefits that we are seeing. ” – Esat Sezer, SVP & CIO

Case Study About Solution § Complex Monte Carlo simulations to calculate financial risks §

Case Study About Solution § Complex Monte Carlo simulations to calculate financial risks § Massively scalable compute leveraging cloud capacity § Daily risk analysis load equivalent to processing 4 trillion US Stocks § Deployed on 2 K Windows Azure instances; 10 K+ in 2010 § Existing on-premise HPC implementation § Windows Azure Worker Role § Windows Azure Blob and Queue § App. Fabric Service Bus Benefits § Elastic scaling to effectively meet customer demands § Turning elements of its peak load infrastructure from a fixed cost to a variable cost

Internet-Scale Application Architecture Presentation • ASP. NET C#, PHP, Java • Distributed in-memory cache

Internet-Scale Application Architecture Presentation • ASP. NET C#, PHP, Java • Distributed in-memory cache Services • • • . NET C#, Java, native code Distributed in-memory cache Asynchronous processes Distributed parallel processes Transient file storage Connectivity • • • Message queues Service orchestrations Identity federation Claims-based access control External services connectivity Storage • • • Relational & transactional data Federated databases Unstructured, de-normalized data Logical partitioning Persistent file & blob storage Encrypted storage SERVICE BUS ACCESS CONTROL WORK FLOWS

Application Patterns Cloud Web Application User Enterprise Application Enterprise Web Svc Web Browser Mobile

Application Patterns Cloud Web Application User Enterprise Application Enterprise Web Svc Web Browser Mobile Browser ASP. NET (Web Role) ASP. NET (Web Role) Private Cloud User Data Application Data WPF Application Jobs (Worker ASP. NET Role) ASP. NET (Web Role) Web Svc ASP. NET (Web Role) (Web Role) Table Storage Service Enterprise Data Enterprise Identity Silverlight Application Reference Data Blob Storage Service Queue Service Bus Application Service Data Service Storage Service Access Control Service Workflow Service Identity Service Public Cloud Services

Application Patterns Composite Services Application User Enterprise Application Enterprise Web Svc Web Browser Mobile

Application Patterns Composite Services Application User Enterprise Application Enterprise Web Svc Web Browser Mobile Browser ASP. NET (Web Role) ASP. NET (Web Role) Private Cloud User Data Application Data WPF Application Jobs (Worker ASP. NET Role) ASP. NET (Web Role) Web Svc ASP. NET (Web Role) (Web Role) Table Storage Service Enterprise Data Enterprise Identity Silverlight Application Reference Data Blob Storage Service Queue Service Bus Application Service Data Service Storage Service Access Control Service Workflow Service Identity Service Public Services

Application Patterns Cloud Agent Application User Enterprise Application Enterprise Web Svc Web Browser Mobile

Application Patterns Cloud Agent Application User Enterprise Application Enterprise Web Svc Web Browser Mobile Browser ASP. NET (Web Role) ASP. NET (Web Role) Private Cloud User Data Application Data WPF Application Jobs (Worker ASP. NET Role) ASP. NET (Web Role) Web Svc ASP. NET (Web Role) (Web Role) Table Storage Service Enterprise Data Enterprise Identity Silverlight Application Reference Data Blob Storage Service Queue Service Bus Application Service Data Service Storage Service Access Control Service Workflow Service Identity Service Public Services

Application Patterns B 2 B Integration Application User Enterprise Application Enterprise Web Svc Web

Application Patterns B 2 B Integration Application User Enterprise Application Enterprise Web Svc Web Browser Mobile Browser ASP. NET (Web Role) ASP. NET (Web Role) Private Cloud User Data Application Data WPF Application Jobs (Worker ASP. NET Role) ASP. NET (Web Role) Web Svc ASP. NET (Web Role) (Web Role) Table Storage Service Enterprise Data Enterprise Identity Silverlight Application Reference Data Blob Storage Service Queue Service Bus Application Service Data Service Storage Service Access Control Service Workflow Service Identity Service Public Services

Application Patterns Grid / Parallel Computing Application User Enterprise Application Enterprise Web Svc Web

Application Patterns Grid / Parallel Computing Application User Enterprise Application Enterprise Web Svc Web Browser Mobile Browser ASP. NET (Web Role) ASP. NET (Web Role) Private Cloud User Data Application Data WPF Application Jobs (Worker ASP. NET Role) ASP. NET (Web Role) Web Svc ASP. NET (Web Role) (Web Role) Table Storage Service Enterprise Data Enterprise Identity Silverlight Application Reference Data Blob Storage Service Queue Service Bus Application Service Data Service Storage Service Access Control Service Workflow Service Identity Service Public Services

Application Patterns Hybrid Enterprise Application User Enterprise Application Enterprise Web Svc Web Browser Mobile

Application Patterns Hybrid Enterprise Application User Enterprise Application Enterprise Web Svc Web Browser Mobile Browser ASP. NET (Web Role) ASP. NET (Web Role) Private Cloud User Data Application Data WPF Application Jobs (Worker ASP. NET Role) ASP. NET (Web Role) Web Svc ASP. NET (Web Role) (Web Role) Table Storage Service Enterprise Data Enterprise Identity Silverlight Application Reference Data Blob Storage Service Queue Service Bus Application Service Data Service Storage Service Access Control Service Workflow Service Identity Service Public Services

Internet-Scale Application • 2007 stats: – +20 petabytes of data processed / day by

Internet-Scale Application • 2007 stats: – +20 petabytes of data processed / day by +100 K Map. Reduce jobs – 1 petabyte sort took ~6 hours on ~4 K servers replicated onto ~48 K disks – +200 GFS clusters, each at 1 -5 K nodes, handling +5 petabytes of storage • ~40 GB/sec aggregate read/write throughput across the cluster • +500 servers for each search query < 500 ms • Scaling the process: – Map. Reduce: parallel processing framework – Big. Table: structured hash database – Google File System: massively scalable distributed storage

Internet-Scale Application • 2009 stats: – – – +200 B pageviews/month >3. 9 T

Internet-Scale Application • 2009 stats: – – – +200 B pageviews/month >3. 9 T feed actions/day +300 M active users >1 B chat mesgs/day 100 M search queries/day >6 B minutes spent/day (ranked #2 on Internet) • Scaling the “relational” data: – Keeps data normalized, randomly distributed, accessed at high volumes – Uses “shared nothing” architecture – +20 B photos, +2 B/month growth – 600, 000 photos served / sec – 25 TB log data / day processed thru Scribe – 120 M queries /sec on memcache

Internet-Scale Application Architecture Design Data & Content § Horizontal scaling § De-normalization § Service-oriented

Internet-Scale Application Architecture Design Data & Content § Horizontal scaling § De-normalization § Service-oriented composition § Logical partitioning § Eventual consistency § Distributed in-memory cache § Fault tolerant (expect failures) § Diverse data storage options (persistent & transient, relational & unstructured, text & binary, read & write, etc. ) Security § Claims-based authentication & access control § Federated identity § Data encryption & key mgmt. Management § Policy-driven automation § Aware of application lifecycles § Handle dynamic data schema and configuration changes Processes § Loosely coupled components § Parallel & distributed processing § Asynchronous distributed communication § Idempotent (handle duplicity) § Isolation (separation of concerns)

The Microsoft Cloud Categories of Services Application Services Software Services Platform Services Infrastructure Services

The Microsoft Cloud Categories of Services Application Services Software Services Platform Services Infrastructure Services

Windows Azure for Application Developers SDK Portal Hosted Service Storage

Windows Azure for Application Developers SDK Portal Hosted Service Storage

Roles

Roles

Load Balanced Instances http: //yourapp. cloudapp. net

Load Balanced Instances http: //yourapp. cloudapp. net

Web vs Worker Role

Web vs Worker Role

Storage

Storage

Blobs

Blobs

Queues

Queues

Development Fabric and Storage

Development Fabric and Storage

Portal

Portal

Hello Cloud

Hello Cloud

Hello Cloud – Web Role

Hello Cloud – Web Role

What We’re Going To Code

What We’re Going To Code

What We’re Going To Code • Step 1 - Create a Cloud Service project

What We’re Going To Code • Step 1 - Create a Cloud Service project with a Web Role • Step 2 - Write our code and test it locally • Step 3 – Deploy to the Cloud

Web Role Demo

Web Role Demo

Portal – Allocated

Portal – Allocated

Portal - Run

Portal - Run

Portal - Initializing

Portal - Initializing

Portal - Busy

Portal - Busy

Portal - Ready

Portal - Ready

Hello Cloud – Blobs and Queues

Hello Cloud – Blobs and Queues

What We’re Going To Code • Step 1 – Configure a Storage Account •

What We’re Going To Code • Step 1 – Configure a Storage Account • Step 2 – Upload a Blob • Step 3 – Enqueue a Message

Blob + Queue demo

Blob + Queue demo

Hello Cloud – Worker Role

Hello Cloud – Worker Role

What We’re Going To Code • Step 1 – Add a Worker Role •

What We’re Going To Code • Step 1 – Add a Worker Role • Step 2 – Dequeue Next Message • Step 3 – Update the Blob referenced in the message • Step 4 – Delete Message from Queue

Worker Role demo

Worker Role demo

Hello Cloud

Hello Cloud

New Project

New Project

Add a Role

Add a Role

Write Some Code

Write Some Code

Run and Debug Locally (Hit F 5)

Run and Debug Locally (Hit F 5)

Deploy to the Cloud

Deploy to the Cloud

Create Your Accounts

Create Your Accounts

Deploy Your Service

Deploy Your Service

And You’re Running

And You’re Running

Summary • Same Development Tools and Experience • Local and Cloud Based Environments •

Summary • Same Development Tools and Experience • Local and Cloud Based Environments • Rapid Cloud Development, Deployment and Maintenance

Under the hood: Windows Azure ‘Compute’

Under the hood: Windows Azure ‘Compute’

Windows Azure Push-button Deployment • Step 1: Allocate nodes – – • • •

Windows Azure Push-button Deployment • Step 1: Allocate nodes – – • • • Across fault domains Across update domains Step 2: Place OS and role images on nodes Step 3: Configure settings Step 4: Start Roles Step 5: Configure loadbalancers Step 6: Maintain desired number of roles – Failed roles automatically restarted – Node failure results in new nodes automatically allocated Allocation across fault and update domains Load-balancers

High-Level Architecture Host Partition Guest Partition Applications Guest OS Server Enterprise Virtualization Stack (VSC)

High-Level Architecture Host Partition Guest Partition Applications Guest OS Server Enterprise Virtualization Stack (VSC) Host OS Server Core Virtualization Stack (VSP) Drivers VMBUS Hypervisor NIC Disk 1 Disk 2 Hardware CPU

Image-Based Deployment Maintenance OS Host Partition Host partition differencing VHD HV-enabled Server Core base

Image-Based Deployment Maintenance OS Host Partition Host partition differencing VHD HV-enabled Server Core base VHD Guest Partition Application VHD App 1 Package App 3 Package App 2 Package Guest partition differencing VHD Server Enterprise base VHD Server Core base VHD

 • Your services are isolated from other services – Can access resources declared

• Your services are isolated from other services – Can access resources declared in model only – Local node resources – temp storage – Network end-points • Isolation using multiple mechanisms • Automatic application of Windows security patches – Rolling OS image upgrades Managed code Restriction of privileges Firewall Virtual Machine IP filtering

Fault Domains • Purpose: Avoid single points of failures • Unit of a failure

Fault Domains • Purpose: Avoid single points of failures • Unit of a failure Fault domains – Examples: Compute node, a rack of machines • System considers fault domains when allocating service roles – Example: Don’t put all roles in same rack • Service owner assigns number required by each role – Example: 10 front-ends, across 2 fault domains Allocation is across fault domains

Update Domains • Purpose: ensure service stays up while updating • Unit of software/configuration

Update Domains • Purpose: ensure service stays up while updating • Unit of software/configuration update Update domains – Example: set of nodes to update • Used when rolling forward or backward • Developer assigns number required by each role – Example: 10 front-ends, across 5 update domains Allocation is across update domains

Platform of Choice

Platform of Choice

http: //www. azure. com Sign up at the Windows Azure Platform developers’ portal Windows

http: //www. azure. com Sign up at the Windows Azure Platform developers’ portal Windows Azure access Developer tools White papers Sample applications Plan pilot applications, proofs of concept, and architectural design sessions with Windows Azure partners

Thank you krishna. kumar@microsoft. com http: //azurepilot. com © 2009 Microsoft Corporation. All rights

Thank you krishna. kumar@microsoft. com http: //azurepilot. com © 2009 Microsoft Corporation. All rights reserved. Microsoft, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U. S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.