Pini Cohen Infrastructure Market 2009

  • Uploaded by: Dr Jimmy Schwarzkopf
  • 0
  • 0
  • April 2020
  • PDF

This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA


Overview

Download & View Pini Cohen Infrastructure Market 2009 as PDF for free.

More details

  • Words: 11,412
  • Pages: 165
Trends in the Israeli Infrastructure Market 2009 Your Text here

Your Text here

Pini Cohen EVP & Senior Analyst Architecture & Infrastructure Strategies pinicohenstki.blogspot.com Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

1

Agenda • General 1 2 3 4 5

– – – – –

SOI - service oriented infrastructure Cloud Computing Open Source and other general issues Major Trends General recommendations

• Technology domains – 7 – 8 – 9 – 10 – 6

Development Enterprise System Management Servers Desktop Storage Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

2

This presentation is an extraction from the complete 200+ slides presentation. Latest version is in the WEB

Some of the slides will be presented quickly as a teaser Thank you for your help in STKI’s research Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Scenario: Adapting in Healthy Ways, Rather Than Fixating

General

Adaptive organizations have a core foundation of values and principles that keep them from chasing their tails every time tactical winds blow Cost Risk and Transformation

Speed

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Quality 4

Agility attributes • Economic agility • Technology agility – Infrastructure – Architecture – Application portfolio • Human resource agility • Operational agility • Process agility • Strategic agility • Organizational, innovational, and learning agility

General

Agility Attributes of the Adaptive Organization Technology Agility

Human Resource Agility

Economic Agility Adaptive IT Organization

Operational Agility

Strategic Agility Organization Innovative Learning Agility

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Process Agility

5

What’s the Problem With Infrastructure?

General

• Too much stuff • Handoffs don’t work • Too much theory or too little practice

• Costs too much • Too slow

• Nothing works together

Application Development

Infrastructure Planning

Integration Testing

Deployment Installation

Operations

The infrastructure as we now it today is very rigid and this causes very poor utilization (40% or less) Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

6

Example of Service Definition • • •

Name: Identity infrastructure service Owner: Bob Smith, architect Description –



Use case – – –



Examples – –



– –



Direct use by application (LDAP) Indirect use via Web server (with attribute passing in headers) Direct use by application (security APIs)

– – – – – – – – – –

+ Scalability (over 500 users, etc.) + Scale incrementally using replicas – Direct application support “Included” in e-business costs



Simple authentication is usually enough Replication to scale (mostly read-only)

Component and service manifest

Pricing –

MSFT Active Directory (NOS file and print) MSFT Passport online service

Principles

Service-level matches – – –



Providing user identity information (attributes), including authentication credentials and related SSO services; also offers Web URL permissions



General

API: LDAP, Web server exits, proprietary Presentation: NA Application server: NA (see Web SSO) Integration: Metadirectory utilities Database: iPlanet Directory Server Server HW/OS: Sun Solaris on SPARC … Storage: EMC SAN Network: NA Security: Netegrity SiteMinder Web SSO Management: Delegated admin, …

Maturity

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

– –

Installed since 2001 with all customer names Used by X, Y, Z apps now 7

Storage Services

General

Disk

Tape

Tier

Enterprise

Midrange

Online Capacity/Arc hival

Automated Capacity

Manual

Design

Monolithic

Modular

Modular

ATL

Rack

SCSI/FC

ATA/SATA

FC

People

1.2 Million+

1.2 Million+

400K+

1 Million+

1 Million+

rpm

10K-15K

10K-15K

7.2K

Seek Time

<6ms

<15ms

<1 sec.

<1 min.

days

Key Environments

Mission Critical, OLTP

Business Critical

Fixed Content, WORM, Archival

BU, Archival, WORM

Archival, BU

Drive Interface Drive/Media Reliability MTBF (Hour)

SCSI/FC Drives

Performance:

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

8

Many dimensions for each service Policy

Architecture

Adaptive Storage Resource

5 Application Optimization 5 Asset Management 5 Budget Management 5 Business Continuity 5 Contract Management 5 Contractor Management 5 Negotiation Management 5 Service Level Agreement Management

Storage Architecture Design 5 5 5 5 5 5

Hardware Support Infrastructure Planning Security Management SW Distribution Facility Planning Security

General

Operations Performance Monitoring

5 5 5 5 5 5 5 5 5 5 5 5

Capacity Management Inventory Management Network Monitoring Performance Management Problem Management Production Control SW Management Service Level Management Workload Monitoring System Monitoring Tape Management Database Adm

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Resource Allocation

5 Change Management 5 Configuration Management 5 Cost-Recovery Management 5 Physical DB Management 5 Disk Storage Management 5 Job Scheduling 5 Service Request Management

Availability Assurance

5 Production Acceptance 5 QA 5 Test Lab Management

9

What Are the Benefits of a Ensemble Infrastructure? •

Technology reuse

Ensemble

– Pattern blueprints: Architecture, technology, product, configuration – Technical services: Actual implementations



Process

IT Process reuse – – – –



General

Pattern matching Service support Predictive costing Experience gained: Good and bad practices

Technology P2 P7 P 1 P5

People

P8

P1 P3

P3 P4 P6

P7 P8

People reuse – Fewer technology skills specific to the pattern – More common roles focused in fewer service areas

J1

T2 T1 T6

T7 T3

T8 T4 T5

J6 J2

J4 J7 J8

T2 T3

J4 J6

J5

Today : Too many technologies, too many processes, not enough people Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

J1 J2

T1

J3

10

Infrastructure Services

General

• Infrastructure services are very appealing • However in many cases this involves: – Saying “no”

– Buying “more expensive equipment” than needed

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

11

Infrastructure Services are the building blocks of Ensembles

General

Ensemble is a combination of specific Infrastructure Services that enable execution of specific Mashup for specific Business Process T1

T6 T3

T2 T7

T5

T8

T4

Gateway T8

T7 T1

Applications Services

T3 T5

T4

T6 T2

T1

T7 T8 T3 T4

T2 T5 T6

Infrastructure Services

Puzzle Builder

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

12

Cloud Computing Modularity

General Cloud

Desktop as a service Software as a service (standard applications) Platform as a service (custom applications) Infrastructure as a service (compute, storage, network) Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

13

Requirements for Cloud Services • • • • • • • •

General Cloud

Multitenant. A cloud service must support multiple, organizationally distant customers. Elasticity. Tenants should be able to negotiate and receive resources/QoS ondemand. Resource Sharing. Ideally, spare cloud resources should be transparently applied when a tenant’s negotiated QoS is insufficient, e.g., due to spikes. Horizontal scaling. It should be possible to add cloud capacity in small increments; this should be transparent to the tenants of the service. Metering. A cloud service must support accounting that reasonably ascribes operational and capital expenditures to each of the tenants of the service. Security. A cloud service should be secure in that tenants are not made vulnerable because of loopholes in the cloud. Availability. A cloud service should be highly available. Operability. A cloud service should be easy to operate, with few operators. Operating costs should scale linearly or better with the capacity of the service. http://209.85.129.132/search?q=cache:IRrnbGDGMLsJ:www.cse.iitb.ac.in/co mad/2008/uploads/K_Raghu_CloudComputing.ppt+Multitenancy+Elasticity+fil etype:ppt&hl=iw&ct=clnk&cd=2&gl=il&client=firefox-a http://www.cse.iitb.ac.in/comad/2008/uploads/K_Raghu_CloudComputing.ppt

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

14

Amazon EC2

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

General Cloud

Source: http://aws.amazon.com/ec2/#pricing

15

Mosso - Cloud Files

General Cloud

1 TB = 4K$~ per year

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

16

Amazon SimpleDB

General Cloud

• web service providing the core database functions of data indexing and querying

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

17

Amazon Simple Queue Service (Amazon SQS)

General Cloud

• Offers a reliable, highly scalable, hosted queue for storing messages as they travel between computers • Pay only for what you use. There is no minimum fee. Estimate your monthly bill using AWS Simple Monthly Calculator. • $0.01 per 10,000 Amazon SQS Requests ($0.000001 per Request) • Amazon SQS requests are CreateQueue, ListQueues, DeleteQueue, SendMessage, ReceiveMessage, DeleteMessage, SetQueueAttributes and GetQueueAttributes • Data Transfer • * $0.100 per GB – all data transfer in • * $0.170 per GB – first 10 TB / month data transfer out • http://aws.amazon.com/sqs/#pricing 18 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

SQL/ACID

Consistency model

Updates

Structured access

Global low latency

Availability

Operability

Elastic

So many new alternatives-players!

General Cloud

Sherpa Y! UDB MySQL Oracle HDFS BigTable Dynamo Cassandra http://www.cse.iitb.ac.in/comad/2008/uploads/K_Ragh Pini Cohen’s work Copyright 2009©STKI 19 Do not remove source or attribution from any graphic or portion of graphic u_CloudComputing.ppt

19

Data Center Locations Netherlands Illinois

General Cloud

China

Ireland Japan

Washington

India

Virginia

California Texas

Microsoft is planning to have 1M servers by the end of 2009

Singapore

Puerto Rico

Current Online locations Proposed future Online locations Other Microsoft locations

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

https://partner.microsoft.com/download/sverige/4 0097312 STKI modifications

20

Microsoft Cloud Services

General Cloud

Azure Services Platform ™

Microsoft SharePoint Services

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Microsoft Dynamics CRM Services

Microsoft: Software + Services

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

General Cloud

The Cast Iron Integration Solution Traditional Approach

Today’s Approach

Benefits

CRM

• No Software to install or maintain • Configuration, not coding • Weeks, not months

Enterprise CRM Software

Integration

General Cloud

Home Grown CRM

Enterprise Integration Home grown Software integration (TIBCO, webMethods) Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: http://salesforce.vo.llnwd.net/o1/us/community/ppt/TEX004 _Cast%20Iron.ppt

23

Flexible Deployment Options

General Cloud

Integration on Premise

Customer’s Data Center

Virtual Appliances Powered By Cast Iron*

Integration as a Service

Cast Iron Cloud™

Source: http://salesforce.vo.llnwd.net/o1/us/community/ppt/TEX004_Cast%20Iron.ppt

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

24

Cloud Computing Challenges

General Cloud

Source: http://www.reservoir-fp7.eu/twiki/pub/Reservoir/PresentationsPage/081126_-_RESERVOIR_Overview_-_TAB.ppt

http://gigaom.com/2008/07/01/10-reasons-enterprises-arent-ready-to-trust-the-cloud/

25

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

25

Gmail issues

General Cloud

http://www.haaretz.co.il/captain/spages/1066642.html Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

26

Example for potential new business model

General - Cloud

Old Fashion

Different SAAS providers (Salseforce Netsuite) + integrator

Integrator is responsible for different SAAS and hosting

Licensing

User

Pay per use – managed separately for each SAAS provider

Pay per use- to the integrator

Servers HW

User owned – Integrator responsibility

At different SAAS provider’s site and in the users location

AT SAAS and integrator site

PC’s

User owned – Integrator responsibility

Integrator responsibility

Integrator responsibility

User perspective

Integrator Split responsibilities between responsibility but Integrator and SAAS providers expensive. Utilization is a joke!

Integrator responsibility

Integrator perspective

Bad -The SAAS provider might add more services that will Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic take the integrator’s place.

Good but costly - needs to maintain high skills for 27 hosting\SAAS

Good

New player – Integration Business & SaaS Solutions(IBSS) • IBSS is new player – providing variety of hosting services and links\integration to different SAAS providers • IBSS clients are the integrators • Example: Spine.co.il

General – Cloud last

With IBSS Licensing

Pay per use- to the integrator

Servers - HW

AT SAAS and IBSS site

PC’s

Integrator responsibility

User perspective

Integrator responsibility

Integrator perspective

Good and cheaper – less skills are needed \ less investment

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

28

Why Companies Use Open Source

General Open Source

• Freedom / Flexibility • Lowers barriers to entry / exit • Can’t afford to build new applications from scratch • Tired of waiting on hold for support • Speed up development • Purchase only the functionality needed • Total Cost of Ownership (TCO) • ROI is shortened

Source: AnyOpen Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

29

Solution - OSS quality • Emerging Vendors Utilize OSS – Netezza • Data warehousing appliance – –

General Open Source

Defects Per Thousands lines of Code

Commodity hardware (servers, CPUs, storage) PostgreSQL, Linux

0.00057

• End user clients utilize OSS – Retail – in-store appliances – Travel – web sites, fee schedules, reservations

0.00009

MySQL

Commercial Source: Reasoning Inc.

Be assured the next generation of developers/entrepreneurs will build on OSS Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

30

General Open Source

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

31

General Open Source

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

32

General Open Source

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

33

Microsoft OpenSource site

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

General Open Source

34

Microsoft IIS: An Interoperable Web Server

General Open Source

• Build and Run Web applications in a high-performance and reliable way along side ASP.NET with IIS 7.0 FastCGI.

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

ISV’s are offering Open Source as alternative for commercial products

General Open Source

• IBI is offering R-Statistics as basic alternative for MATLAB

http://www.informationbuilders.de/Kundentage-presentations/15-10-08/3_IBI_BI_2-0_DaveSmall.pdf Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

STKI Madad

General Open Source - last

•Lately we have received lots of Open Source related inquiries – maturity, support, reference, prices of: •OpenOffice •Tomcat •Jboss •MySQL •Drupal •Nagios •Subversion

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

37

Online Bids

General Procurement

• Online bids are powerful tool in case of real competition of commodity products (apples vs. apples) • Suggested steps are – technical à traditional negotiation for getting 2-3 suppliers for best of finalà online bid • In case there is no real competition online bids can not help

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

38

Virtualisation Licensing Complications

General Procurement

• Historic

compatibility (product & license – “serial machine number”) • PUR’s (product user rights) not supporting new technology • Inventory / Discovery tools not designed to support Virtualised environments • Ever more complicated product variations to accommodate model • More software products required to run and support it • Increased licensing knowledge required • Makes planning and budgeting more complicated

STKI Round Table for IT procurement managers at 22.4.09 http://www.socitm.gov.uk/NR/rdonlyres/684C81F1-3327-472C-842568E8F39F79D0/0/S20080912Agenda5AlexLemonCIVICAVirtuallyLicensed.ppt STKI modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

39

Virtualisation Licensing Complications Software ISV is installed server a: 2 cpu dual core

Software ISV is installed server b: 2 cpu dual core

General Procurement

Software ISV is installed server c: 2 cpu dual core

Virtual server a with ISV installed

Virtual server a with ISV installed

Server Z : 4 CPU QUAD CORE

Server X : 2 CPU dual core

STKI Round Table for IT procurement managers at 22.4.09 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Virtualisation Licensing Complications Software ISV is installed server a: 2 cpu dual core

Software ISV is installed server b: 2 cpu dual core

Lower cost? Virtual server c with ISV installed Virtual server b with ISV installed Virtual server a with ISV installed

Server X : 2 CPU dual core

Virtual server c with ISV installed Virtual server b with ISV installed Virtual server a with ISV installed

Server Z: 4 CPU QUAD CORE

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

General Procurement

Software ISV is installed server c: 2 cpu dual core

HP EDS merge

General

Shared Standard Services Model and expertise for mid-market customers

Established, marketleading services … breadth and depth Industry reach across major industries

High outsourcing market growth rates and momentum

Global Delivery

Innovation from HP Labs, annual R&D investment of $3.6B2

• Enterprise Service Mgmt.

42

• EDS Best Shore® locations

Market leadership across ITO segments

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Improved Efficiency of IT Service Delivery

Technology platform leadership

Comprehensive end-to-end IT services Portfolio Deep industry Expertise, delivered Globally Based on market-leading technology and Innovation

Increased business value from IT

Maximum Client Value

42

New business models\ alternatives

We acquire BEST solutions from LEADING ITs

General

We sell those solutions to other ITs

www.bits.co.il Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Recommendations

General Recommendations

4Less technologies to take care of. Can existing technologies \ products do the job? 4Define Infrastructure Services 4New technology ROI is tricky – the new technology can save money but it requires new processes, people, integration Your Text here Your Text here 4Open Source 4IaaS for: • • • •

Peak time activities Backup Testing applications and new technologies Training and Development Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

44

Recommendations :

General Recommendations

4SAAS for internal IT needs –Example ALM tools as Saas 4SBC (traditional & VDI) for remote locations , call centers 4Operations Scorecards for better alignment 4Datacenter consolidation, Server Consolidation (to Your Text here Your Text here standard platforms) and Server Virtualization 4Asset management in general and usage\metering 4Storage read\write “snap on snap” for cost efficient environment creation 4Vendor and Contract management (STKI can help!) Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

45

General

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

46 46

STKI IT Model

Development and Middleware Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

47

Technologies Maturity Model

2009 – Middleware and Development

DEV Trends

Business Value

Investment to make money

AGILE

BPM GUI integration

Cut costs, Increase productivity

WPF

TDD

Semantic

Full SOA – Organization change

Open Source ALM tools

Commodity IT Services

EAI ETL

SOA Governance tools

PaaS

IT Project Pure Business Project

Investment for regulations Using

Implementing

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Looking 48

New Generation of Knowledge Workers

DEV Trends

• • • •

Live in virtual worlds Expect immediate results Many open threads Deliver quickly, in small increments • Expect personalization • Intolerant of disruption In short: very agile! http://www.solutionsiq.com/resources/SIQAgileDevelopmentPlatform-Rudd-YoungAgile2008.ppt Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

49

Platform convergence

DEV Trends

• “Developers are customers too” – Demand the same features as the intended users (multi-medial, social networking, personalization, customization, etc.)

• Developers should live in the same world they are constructing • Very Agile Source: http://www.solutionsiq.com/resources/SIQAgileDevelopmentPlatform-Rudd-Young-Agile2008.ppt STKI modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

50

Agile Project Management Overview

DEV Trends

Implement

Test Test

Code

Code

9

Analysis Analysis

12 6 3

3

6

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

9

12

Source: Cutter

51

Cost of Feedback – Test Driven Development

DEV Trends

http://losangeles.pm.org/presentations/tdd/TheRoadToTestDrivenDevelopment.ppt

52 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

52

Test-Driven Development • • • • • • •

DEV Trends

Pick a Feature Write a Test for that Feature Run all Tests and see new Test Fail Write Code for Feature Run all Tests and see all Tests Pass Refactor Repeat http://losangeles.pm.org/presentations/tdd/TheRoa dToTestDrivenDevelopment.ppt Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Copyright Oversee.net 2008

53

Composite Application Platform in the new Open Source – Cloud world Development and Test Tools

RIA RIA

Collaboration: Wiki, Blog, Chat

Custom Productivity Tools - Conferencing

Social Networking: People Activities Associations

Business Intelligence: Reporting Analysis Search

Enterprise Content Management: Management: Document Repository Versioning Web publishing Workflow Security and Authentication Authentication

Integrations with existing networks and Integrations existing networks functionalitywith – Mashups, OpenID

Plug-ins and Extentions

Web Portal

DEV Trends

Hosted SaaS Deployment Environment

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://www.solutionsiq.com/resources/SIQ54 AgileDevelopmentPlatform-Rudd-Young-Agile2008.ppt

Open source solutions for .NET environment

DEV Trends

• Spring open source programming framework also for .net! • Nhibernate (Object/Relational Mapping ): – ORM library implemented using .NET – Port of the Java library called Hibernate – Database-independent (almost) – Provides most advanced features – Free, open source and mature http://webseminar1.xmp02.host.35.com/webseminar/NHibernate_in _Action_Seminar.ppt stki modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

55

RIA - Rich Internet applications • Rich Internet applications (RIAs) are web applications that have some of the characteristics of desktop applications, typically delivered by – web browser plug-ins, javascript compiler , etc. – independently via sandboxes or virtual machines

• Examples are : Adobe AIR, JavaFX, Microsoft Silverlight, Magic’s UniPaaS Source : http://en.wikipedia.org/wiki/Rich_Internet_application stki modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

56

Mashups - Where Can I Live?

DEV Trends

• Find a new place to live with a travel time search. • It works on commuting time between stations (Underground and DLR only) and average house prices. You can also view property listings around your chosen station and fun stuff like house price statistics. • APIs CloudMade + Google Maps + Nestoria + OpenStreetMap From http://www.programmableweb.com Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

57

http://www.where-can-ilive.com/londonproperty

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

DEV Trends

58

DEV Trends

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Multicore programming skills

DEV Trends

60

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://www.infoworld.com/article/08/03/20/Multi core-boom-needs-new-developer-skills_1.html



60

MDM – Master Data Management • Typical workflow

• New workflow

Search of customer exist – in the current application domain

Enter details of new customer

DEV Trends

Search of customer– in MDM

Enter details of new customer in MDM

Enter details of transaction sometimes based on insights taken from MDM Enter details of transaction Sometimes if online search is not possible -add new phase to process – “end of day” data corrections

End END Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Enabling MDM in new applications • Application Architecture: – Data of MDM candidates should be kept separately – Function for identifying existing customers

• Different business process (users): – Identifying the customer is part of the process – Sometimes the identification is done after the customer is left

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

DEV Trends

Local Trends - Platforms

DEV Local Trends

• .Net is very strong in the general IT development scene • Many clients are suspicious towards Open Source. There are some innovators (Harel Insurance, Direct Insurance, etc.) • Users are still using legacy development environments – Cobol , Natural, PowerBuilder • Magic uniPaaS RIA & SaaS: New Rich internet & SaaS application platform based on .NET Technology (Launched on January 2009) Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

63

Local Trends - Estimation

DEV Local Trends

• Israeli project manager do not use formal estimation methodologies for estimation (function points, cocomo, etc.) • Each member of project gives estimation “to the best of his knowledge” • Some organizations have estimation sheets with questions like “does the project needs information from the DW?” • Only few organizations track the real usage of specific requests (after went to production) Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

64

Local Trends - SOA

DEV Local Trends

• Most users have EAI tools. Some SOA tools. • Run Time SOA governance (mainly monitoring) is an issue to many users • Dev. SOA governance tools are considered as part of general SOA adoption • What stops SOA are related organizations issues – development roles are different – the architect is in the center • Sometimes the EAI\SOA team creating bottleneck for the projects • STKI Round Table: What are the optimal technology and processes that enable code and interfaces\services reuse? At 7.6.09 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

65

Local Trends - Agile

DEV Local Trends

• Many organizations did not try “agile software development” formally • Many organization have tight relationship with their users – try to deliver in short release cycles, however official project design is done • “Agile Sparks” offers training and accompanying scrum processes mainly to ISV’s. The “end user” is the marketing manager • Some nice first steps for Agile in Israeli enterprise (STKI Round Table) – More details at http://www.pdfcoke.com/doc/12901045/Agile-Round-Table Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

66

The PMO-Development gap?

DEV Local Trends

• What is PMO? – – – – –

IT Plan Development Demand Management Budget and Resource Management Post implementation audits Managing large\strategic\cross LOB projects

• What about PMO and Dev methodologies\technologies ? – – – –

SOA Reuse Agile TDD Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

67

DEV Local Trends

EAI\SOA Support Ratios • EAI \ SOA variations are very big: – – – – –

From static (old) messaging interface (MQ) to new and changing “SOA” service How “thick” is the interface: 20 times a day to 20,000 a minute How involved are the applications programmers ? How static are the interfaces? How many new interfaces? Does “one user” for web service is valid “Services count”?

• EAI\SOA typical Support Ratios : from 10 interfaces\services per FTE to 300 interfaces\services\web services per FTE . Median is 40 interfaces. • This demonstrates the large variety between organization with their perspective to EAI\SOA\Integration

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: STKI 68

STKI recommendation

DEV Local Trends

• Can I define Service\Standards for Development \ Middleware ?

• STKI Round Table: What are the optimal technology and processes that enable code and interfaces\services reuse? At 7.6.09 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

69

STKI IT Model

Enterprise System Management

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

70

Technologies Maturity Model

2009 – Enterprises System Management

ESM Trends

Business Value

Investment to make money Cut costs, Increase productivity

Application Monitoring

Commodity IT Services

End User Experience

System Management

IT Project

RBA

BSM

CMDB – Basic – Automatic discovery

Pure Business Project

CMDB with closed loop change management

Investment for regulations Using

Implementing

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Looking 71

Definition of the Mgmt Segments System & Config Mgmt

Service Management

Performance & Monitoring

Operations Automation

ESM Trends

Technologies used to understand and manage what resources an organization has as well as how, when, and who alters it all Covers all aspects of software, hardware, and licenses/contracts Technologies used to control how an organization delivers and manages IT services supplied to the business Act as the primary communication point between IT and the rest of the organization Technologies focused on observing and ensuring the health and availability of infrastructure Used to understand the current state of all hardware, software, and business services Technologies used to automate day to day operational tasks Used to assist IT staff in automation of routine tasks

ESM - Service Desk – Asset – System (configuration) – Security events Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

72

CMDB automatic discovery

ESM Trends

Learn how their CIs are configured (& changing over time) Automatically tracks changes on all CIs & attribute values over time…

Configuration Auditing – Tracks changes in applications – Depicts that information on the map – Depicts that information thru reports

73

Application

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: IBM

73

CMDB is improving configuration management

ESM Trends

Comparing two instances of an Apache Web Server to the golden master Compliance – Compare configuration to “reference master” – Compare to your standard policy

74

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Values in red and blue are policy violations

Source: IBM

74

Local Trends – change management

ESM Local Trends

• One week ago we have switched off backup communication line but found out that it was used by application in production" • "Expect the unknown results when installing new patch" • Number of changes in large bank (MF environment) 80K a year! • Above 50% of problems are related to changes Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

75

Local trends

ESM Local Trends

• Who is the ESM “owner” ? Different approaches: • “High Touch” ESM team: – The different system teams (Win, Unix, Network, PC) install the agents but with strict instructions from the ESM team – Single console maintained by the ESM team

• The ESM team is in the “background” – The different teams have their own management consoles (example – Win- MOM, Unix- BMC, Network – NMM, PC – Aternity) – Manger of managers console is updated by the ESM team

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

76

ESM (Enterprise System Management) Ratios

ESM Local Trends

• ESM Ratios are very different from companies implementing ESM solutions since the level of implementation, internal processes , tools and automation is very different. • The KPI itself –” per Servers “– is not ideal as well since it doe not give indication to Routers, Storage devices, etc. • From 110 Servers monitored by 1 FTE to 500 Servers monitored by 1 FTE. .Median is 300 Servers • Increase of about 20% from last years data • Data point in the STKI survey 1 FTE per 1000 server. STKI comment “close the project!” Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: STKI

77

STKI recommendation

ESM Local Trends

• Can I define Service\Standards for System Management \ Monitoring?

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

78

STKI IT Model

DBMS Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

79

DBMS Trends

The evolution of distributed architecture

G U I

Bottleneck's…

DB

D A L

B L

B L

U U

G U I

U

G U I

U

SCALE

Scale up by working with layers, Try not to bother the database as much as possible!

U

U

SCALE Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://download.microsoft.com/download/A/2/0/A20CB03B80 7F99-4A16-9652-9A2DAC44986F/ARC301.ppt

What is Velocity ?

DBMS Trends

An explicit, distributed, in-memory application cache for all kinds of data (CLR objects, rows, XML, Binary data etc.) Fuse "memory" across machines into a unified cache Clients can be spread across machines or processes

Unified Cache View

Clients Access the Cache as if it was a large single cache Cache Layer distributes data across the various cache nodes

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://download.microsoft.com/download/A/2/0/A20CB 03B-7F99-4A16-9652-9A2DAC44986F/ARC301.ppt

Semantic Technology

DBMS Trends

Semantic technology as a software technology allows the meaning of information to be known and processed at execution time. For a semantic technology there must be a knowledge model of some part of the world that is used by one or more applications at execution time

Source: Vagan Terziyan

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

82

SPARQL Query: Access Multiple Sources PREFIX dc: PREFIX ldap: PREFIX foaf:

DBMS Trends



SELECT ?name ?name2 { ?doc dc:title ?title . FILTER regex(?title, “SPARQL”) . ?doc dc:creator ?reseacher . ?researcher ldap:email ?email . ?researcher ldap:name ?name }



“Find the name and email addresses of authors of a paper”



Query tech reports first, then go to LDAP for more information.



Other examples: •Copy

schedule of conference to Outlook calendar.

•Copy

list of people to Outlook contacts

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: HP

STKI modification

83

DBPedia

DBMS Trends

. Extracting Structured Information from Wikipedia

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://ids.snu.ac.kr/w/images/3/3c/SC14.p pt dbpedia STKI modifications

84

SPARQL Query PREFIX skos: SELECT ?presName WHERE { ?presName skos:subject . }

From http://dbpedia.org/sparql

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

DBMS Trends

Increase efficiency: Knowledge Pack examples Ø

DBMS Trends

When (identification): (1) SQL requests run full scan and not through an index (2) SQL requests scan more then 100 daily partitions on selected large tables,or (3) SQL requests do not include a restricting date condition, or (4) DML/DDL/DCL commands issued by unauthorized user

Ø

Then (one or more from these actions): (1) Apply a rewrite/adding an Oracle hint (2) block and return a message to the user, to run after hours (3) automatically switch to run on replication

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

86 86

Increase efficiency: MoreVRP (Virtual Resource Partitioning)

DBMS Trends

• Define computing resources (CPU, I/O) for specific transactions In real-time, based on the business needs and Qos requested. • Improve the performance of your current hardware and software in 20-40%"

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

87

Increase efficiency: Exadata – A New Architecture

DBMS Trends

• Exadata Ships Less Data Through Pipes – Query processing is moved into storage to dramatically reduce data sent to servers while offloading server CPUs

• Exadata has More Pipes – Modular storage “cell” building blocks organized into Massively Parallel Grid – Bandwidth scales with capacity

• Exadata has Bigger Pipes – InfiniBand interconnect transfers data 5x faster than Fibre Channel

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Exadata Moves a Lot Less Data a Lot Faster 88

Increase efficiency: SQL Server® Fast Track Data Warehouse • New Data Warehouse reference architectures • Based on SQL Server 2008 + optimized commodity hardware from HP and Dell • Advantages: vAccelerated time to solution vReduced risk – pre-tested; application of best practices; optimized hardware vScale predictably to 30 TB without extensive tuning vLow TCO/performance through commodity hardware

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

89

Other companies for increased performance \ utilization

DBMS Trends

• EdenBase -In the context of “Pini Cohen“ in stead of 0547000023 (10,000 times) write XX

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://www.edenbase.com/Edenbase_CompactDB_Whitepaper.pdf

DBMS Local Trends

DBMS Support Ratios • DBMS support ratios metrics are not trivial: – Per DBMS but sometimes there are several DBMS single application or one DBMS for several applications – Huge variety. In the same organization: • 20 application per FTE for MSSQL (small applications) • 1.5 applications per FTE for Oracle (core business)

– Applications DBMS vs. Infrastructure DBMS – Sometimes part of the job is done internally and some by the vendor (application producer)

• Infrastructure DBA Support Ratios (all copies of application, Open) : from 3 applications per DBA FTE to 50 applications per DBA FTE. • Median is 10 applications per DBA FTE Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: STKI 91

STKI recommendation

DBMS Local Trends

• Can I define Service\Standards for DBMS ?

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

92

STKI IT Model

Servers Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

93

Technologies Maturity Model 2009 - Servers

Server Local Trends

Business Value Investment to make money Cut costs, Increase productivity

Commodity IT Services

Server Virtualization for Prod Linux Servers

Server Unix\MF Cloud \Legacy consolidation Servers

IT Project Pure Business Project

Win 2008 Server

Investment for regulations

Using Implementing Looking Mature technologies. Too many servers. Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

94

Open Virtual Machine Format (OVF)

Server Trends

A multi-vendor format VM 1

License

VM 2

...

VM 3

Properties

import http://lycos.dropcode.net/VMware/Documentations/vmworld2008/ TA10.ppt Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

95

VMware vCloud: Extended DataCenter DataCenter

Virtual Center

Server Trends

Corporate or Public Cloud Based Service

API

Remote Monitoring of VM based applications

Policies Move with VM based application

V-Motion Simplified migration of VM based application Current

Underway

http://www.irislink.com/Partners/irislink2009/PP T/IRISLink2009_MO_Danny%20Roefflaer.ppt Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

96

Transparency in the Eye of the Beholder

Server Trends

…but its difficult to correlate network and storage back to virtual machines

http://communities.vmware.com/servlet/JiveServlet/previewBody/9523-102-1-6774/Nexus-1000V-VMwareUG.ppt;jsessionid=880F4798E7D5D028FE6A3F9EF9111439 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

97

Cisco Virtual Network Link – VN-Link Virtualizing the Network Domain

Server Trends

Problems: VMotion

VLAN 101

• VMotion may move VMs across physical ports—policy must follow • Impossible to view or apply policy to locally switched traffic • Cannot correlate traffic on physical links—from multiple VMs VN-Link: •Extends network to the VM •Consistent services •Coordinated, coherent management

http://communities.vmware.com/servlet/JiveServlet/previewBody/9523-102-1-6774/Nexus-1000V-VMwareUG.ppt;jsessionid=880F4798E7D5D028FE6A3F9EF9111439 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

98

Windows Server Datacenter Virtual Sessions Per License

Unlimited .. 32 16 8

4

1 Host OS Included Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Server Trends

What’s Next For Xeon-EP? NEHALEM NEHALEM

Up to 25.6 Gb/sec bandwidth per link

Nehalem

Key Technologies

Nehalem

QPI

I/O Hub PCI Express* Gen 1, 2

DMI ICH

Server Trends

Functional system demonstrated Sept 2007 IDF

• • • • •

New 45nm Intel® Microarchitecture New Intel® QuickPath Interconnect Integrated Memory Controller Next Generation Memory (DDR3) PCI Express Gen 2

IT Benefits • More application performance • Improved energy efficiency • End to end HW assist for Virtualization • Stable IT image

Extending Today’s Leadership Launch Q1’09 Volume Ramp 1H’09 https://www.imsc.com/scevents/IBM/IBM%20Partner%20Day%20Documen ts/Intel%20Update%20for%20Ingram%20BP%20110608.ppt

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Server Trends

Technology Investment Areas - 2010 Virtualization

Web

Management

Hyper-V™

Power Management

IIS 7.5

Live Migration

PowerShell 2.0

.NET Support on Server Core

Cluster Shared Volumes

AD Administrative Center

Integrated IIS Extensions

Remote Desktop Services

New Best Practices Analyzers

Configuration Tracing

Solid Foundation for Enterprise Workloads Scalability and Reliability 64+ Logical Processor (LP) Support Componentization DHCP Failover DNSSEC

Better Together with Windows 7 DirectAccess BranchCache™ Enhanced Group Policies BitLocker on Removable Drives RemoteApp & Desktop Connections

64-bit only Server release

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

RHEV - HYPERVISOR Scalability Host: 96 cores, 1 TB RAM Guest: 16vCPU, 64 GB RAM

Industry Standards Trusted RHEL kernel + KVM High performance VirtIO drivers Libvirt management interface

Small footprint – easy to install and manage Boot via PXE Run from flash drive Install on local media

Advanced Features Memory Page sharing SELinux for high security and isolation Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Server Trends

Automation is key enabler

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Server Trends

103

Server market is down – how low can we get?

Server Local Trends

4Recent data shows that the Israeli server market in 2009 is about 40% lower than last year 4Will this continue?! 4Who will be caught in the fire?! Your Text here

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Your Text here

104

Local Trends – Server Virtualization

Server Local Trends

• Most servers sold \installed within the enterprise IT organizations are virtual servers • Users might install application in Virtualized environment even without ISV permission! • Percent of Virtualized Windows servers is between 5% and 40% (some extreme cases more). All clients want to increase this ratio. Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

105

Local Trends -Server Virtualization

Server Local Trends

• Server virtualization and Outsourcing – server provisioning in virtualized environment – 2 hours before it was 10 hours • Issues with virtualization: – – – –

Backup – agents, copy of vmdk’s, vcb – vmware consolidated backup DRP\HA is in the LUN level (many vm guests) not for particular guest Network virtualization Cloud offerings

• Microsoft is investing a lot in Server Virtualization however the integrators do not get revenues from hyperV projects and this makes things a bit difficult for Microsoft Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

106

IT organization - What is Infrastructure?

Server Local Trends

Networks, Voice Storage , Backup, DRP System – Legacy, Windows, Unix, Linux Data Center operations and facilities PC – preparing the images, first level support (remote control) , second level support • DBMS support – DBA’s • Integration – SOA - MQ • System Management – configuration the ESM tools

• • • • •

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

107

Org. Chart A

Server Local Trends

Infra

Telephony

Unix Storage

Windows Servers and PC + storage

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Operations, Datacenter , Service desk

108

Org. Chart B

Server Local Trends

ESM (engineering)

Service desk Infra

dba

security

Network

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Datacenter operations and PC

System , storage sap basis, integration

109

Server Local Trends

Org. Chart C Infra

legacy

System open

DBA

Web Infrastructure Networking

Your Text here

Your Text here Operations PC and Datacenter Server Network HW Operations (open) HW

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Service Desk

110

Server Local Trends

Org. Chart D

Infra

Service Desk

Your Text here

Networking cellular video conference

Datacenter operations

Legacy

Your Text here

PC

Security

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

System division

SAP Basis

DBA

111

System and Storage

Org. Chart E

Server Local Trends

Infra

Branches

Security, Networking ESM

Operations: pc second level support

System: Legacy, Unix, MSFT, DBA, Storage, manager of changes 3rd level support including

Service desk –first level support

datacenter

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

112

Org. Chart F

Server Local Trends

Operations: QA, Service desk, datacenter operations Infra DBMS

Network, desktop + 2nd level, datacenter facilities, drp

Legacy Unix general Storage ESM, service desk (application) drp

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Windows servers including backup , security

113

What is good Infrastructure organization?

Server Local Trends

• Possible Alarms for organization issues: – System is using ESM product X. Service Desk uses Y for monitoring the same infrastructure … – There are several different change management processes within the organization – “I know storage product X is more appropriate but I should purchase Y because it will upset one storage team while the other storage team will be happy”

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

114

What is good Infrastructure organization?

Server Local Trends

• STKI recommendations: – Focus on Infrastructure Services – Infrastructure services should match (as much as possible) the org. structure. – Not so politically correct recommendation – Now

is the time for organization changes

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

115

Server Local Trends

Server Ratios - Windows • Number of Windows Servers (logical ) per System member

– In Israel from 20 per FTE in rather small but geographically distributed organization to 200 per FTE in large well managed organization. Median is 100 Windows Servers (logically) per FTE – Change from last year – about 10% increase • Organizations with 100% identical servers in branches can get ratios of 1:500 servers per FTE • The SBC system staff is part of the system or PC count?! Source: STKI Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

116

Server Local Trends

Server Ratios - Windows • Windows Server growth is huge

• Many clients claimed they are understaffed – not able to perform the patches on time, capacity planning, consolidation project is too slow, keeping track of all changes, etc. • STKI observation – Virtualization has potential of savings but must start with adequate resources (staff).

Source: STKI Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

117

Server Local Trends

Server Ratios –Unix Linux

• Number of Unix servers (OS) per System member in Israel is from 6 servers per FTE (rather small, geographically distributed organization) to 80 servers per FTE in large homogeny Telco organization . Median is 27 servers per FTE • Roughly same ratios as last year’s data • Unix is rather stable \ declining . Many clients are migrating to Linux with the same staff • Good metric for Unix is hard to find: – Per CPU (but there are machines with many virtual OS on each CPU) – Per OS (but there are sometimes huge machines with 1 OS) – Per physical server Source: STKI Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

118

Market Status and Recommendations

Server Local Trends

• Users recommended the following server integrators/support :

•HP • • • •

TEAM\Malam IBM EMET WE PenguinIT

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

119

STKI recommendation

Server Local Trends

• Can I define Service\Standards for PlatformsServers?

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

120

STKI IT Model

Desktops Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

121

Technologies Maturity Model 2009 – End Point Devices

PC Trends

Business Value

Investment to make money Cut costs, Increase productivity

Desktop as a Service

Mobile Devices Traditional SBC for niche environment

Commodity IT Services

MAC for the Business

VDI Office 2007

Application virtualization

IT Project Pure Business Project

Vista

Windows 7

Software Appliances

Investment for regulations Using

Implementing

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Looking 122

General Purpose computation on the GPU (graphics processing unit)

PC Trends

– Started in computer graphics community – Mapping computation problems to graphics rendering pipeline

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Courtesy Jens Krueger and Aaron Lefohn Source : http://www.na-mic.org/Wiki/images/5/5c/Hpc-g80-cuda.ppt stki modifications

123

Why GPU for computing?

PC Trends

• Inexpensive supercomputer – Two NVIDIA Tesla D870 : 1 TFLOPS

• GPU hardware performance increases faster than CPU

GFLOPS

– Trend : simple, scalable architecture, interaction of clock speed, cache, memory (bandwidth)

G80GL = Quadro FX 5600 G80 = GeForce 8800 GTX G71 = GeForce 7900 GTX G70 = GeForce 7800 GTX NV40 = GeForce 6800 Ultra NV35 = GeForce FX 5950 Ultra NV30 = GeForce FX 5800

Courtesy NVIDIA Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source : http://www.na-mic.org/Wiki/images/5/5c/Hpc-g80-cuda.ppt stki modifications

124

General Purpose computation on the GPU (graphics processing unit)

PC Trends

• Substantial enhancements in data manipulation and presentation • Standard PCs will greatly expand the types and amount of processing available to the individual user. • It will take some time for SW to catch up with the new HW, but once it does we can expect a substantial increase in the personal productivity and data manipulation and presentation capabilities of the average PC user • Both Thick and Thin Clients will be Thick in CPU power! Source: www.jgoldassociates.com stki modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

125

I’m hungry! Can I get an Apple?

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

PC Trends

Data from : http://arstechnica.com/a pple/news/2009/02/mac -market-share.ars

126

PC Trends

Netbooks •

Potential users will include both first-time buyers seeking a low-cost introductory PC as well as experienced users seeking a low-cost second PC for themselves: • • • •

Communication such as E-mail, VoIP, Video Phone, etc Internet access Content consumption such as music & movie playback Content management such as music & movie download, digital photo management, file & content search, etc CHILDREN • •

Buyer • Usage • • •

Parents/School



Internet access E-Mail Content consumption & management



FIRST TIME USERS • •

TEACHERS

Buyer • Usage • • •

Buyer • School Usage • Internet access • Content consumption & management

SECONDARY PC & LIGHT PC USER •

User • PC Familiarisation Internet Access E-Mail, Entertainment

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Buyer • Usage • • •

User Internet access Document viewing Entertainment

Source: http://www.bms.ie/toshiba/NB100%20Netbook%20Briefing.ppt stki modifications

Windows 7: Core Technology

PC Trends

• Vista Kernel = Better Compatibility • Focus on reliability and performance – Less memory footprint – Greatly improved performance – Less power consumption = “Greener” OS and longer battery life

Fundamentals http://stlouisdayofdotnet.com/FlashDriveContents%5COverview %20of%20PDC/PDC%20Overview%20%20Kevin%20Grossnicklaus.ppt Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

128

Why software can be painful • • • • • • •

PC Trends

Assembly required! User has to acquire all of the required components of one of the “supported” stacks User has to install, configure and maintain everything Manually intensive process Error prone Difficult for solution provider to replicate and support Fault tolerance (HA/DR) not always there, implemented in inconsistent fashion (requiring run books for each process), and/or an expensive add-on Actual Solution Database Software Drivers OS Patches Operating System

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://marketopsrepo.vmware.com/usergroup/preso /Virtual.Appliances.Overview.v3.ppt 129

What is a virtual appliance?

PC Trends

• Pre-installed and preconfigured application packaged along with an OS in a virtual machine • Virtual appliances represent an evolutionary step in the distribution of software solutions

• Updates (application – OS – other) is handballed by the ISV! http://marketopsrepo.vmware.com/usergroup/preso/Virtual.Applia nces.Overview.v3.ppt STKI modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

130

Next Generation Laptops: Virtualization “Built-in”

PC Trends

Will this lead to BYOC personal desktop

corporate image



Bring Your Own Computer?

Hypervisor from Citrix/Intel Intel Processors

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: Citrix STKI modifications

131

Solid State Drives (SSD) for laptops

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

PC Trends

132

Desktop vs. Laptops?

PC Trends

http://www.geekwithlaptop.com/notebook-laptop-salesovertake-desktops-for-first-time/ Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

133

PC Deployment Options (and mix of these options)

PC Trends

PC Deployment

Blade PC

Server Based Computing

Client Technology

Thin Client no OS

Thin OS – WinCE, Linux

Application Streaming

Traditional Deployment

Server Technology

OLD , WinXPE PC

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Citrix, Terminal Server, Jetro

VDI 134

The ClearCube Solution – PC Blades

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

PC Trends

Source: http://www.ehcca.com/presentations/HIPAA9/3_06.ppt 135

Thin Clients

PC Trends

According to a recent tracker study on the Asia/Pacific excluding Japan (APEJ) thin client market, total sales of thin clients in 1H 2007 reached 282,667 units, representing an increase of 37.3% Across the various verticals, the predominant role of thin clients across the region has shifted away from government/education segment to financial services …

Source: http://www.dc.uba.ar/events/eci/2008/courses/n2/VirtualizationFuture.ppt STKI modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

136

PC Trends

VDI – Virtual Desktop Infrastructure

Streaming

• PC Virtualization will ease the burden of new PC-HW assimilation by the organization –the PC HW layer is virtualized = standard! Source: Citrix STKI modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

137

VDI – Virtual Desktop Infrastructure who is doing the rendering?!

PC Trends

Guest Guest

Not using the “real HW”

Host `

The Host – “real HW” but for many users

Thin Client/Desktop Windows/Linux based The client SPICE is an adaptive remote rendering solution Utilizes GPU capabilities of the server and/or client Offloads graphics processing/rendering to client Source : Redhat stki modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Local trends –PC

PC Local Trends

• For XP users (most of enterprises) used PC’s – Pentium 4 , 2G RAM, CD, Speakers, warranty for 3 years at about 700 NIS is an option! • Some users are still buying “IBM Laptops” and not Lenovo

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

139

Local Trends- SBC

PC Local Trends

• SBC and thin clients are mainly used in segments • Some organizations put mandatory for every new application – SBC compatible • ROI example for SBC project – in a call center environment , login in the morning took 15 minutes (“boot storm”). In SBC the session is on instantly. • Less support is needed and uptime has increased sometimes dramatically up to 10 time less incidents!

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

140

Local Trends – VDI

PC Local Trends

• Users are looking carefully at VDI – new form of SBC – Not all application are compatible to traditional SBC – Smaller shock for the end user – still have the same environment

• VDI is maturing – Scalability compared to traditional SBC – SW distribution – Storage needed

• Users look mainly at VMWARE and CITRIX. Also Microsoft and Sun Microsystems

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

141

Local Trends – Vista and Office 2007

PC Local Trends

• Users that migrate to Vista are happy – stability, security, GREEN, etc. • Many XP users will skip Vista and go to Windows 7 • Office 2007 is deployed gradually. Stages are: – Application compatibility and image preparation (around 50K NIS) – Training and then 0.5 days tight support (1400 NIS at clients location per day) Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

142

PC Support Ratios and TCO

PC Local Trends

• PC First Level Support Ratios: from 120 users per Help Desk FTE (rather autonomic organization) to 1400 users per Help Desk FTE in centralized (from image point of view) organization. Median is 360 users per Help Desk FTE • PC second level support ratios: from 175 users per PC technician to 2000 users per PC technician. Median is 400 users per PC technician • System PC team (building the image, third level support for system PC problems) ratios: from 350 PC’s per system PC FTE, to 4000 PC’s per system PC FTE (very stable organization), Median is 1600 PC’s per FTE Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: STKI 143

STKI recommendation

PC Local Trends

• Can I define Service\Standards for End Point Devices – PC’s , Laptops, etc.?

• STKI Round Table: VDI, Traditional SBC, Application Virtualizations at 20.5.09 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

144

STKI IT Model

Storage Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

145

Technologies Maturity Model 2009 - Storage

Storage Trends

Business Value Investment to make money

CDP

Cut costs, Increase productivity

Green Storage

Commodity IT Services Investment for regulations

NAS and SAN

Dedup for Backup

Thin Provisioning Flash Drives

Email Archiving

Using

Implementing

Sophisticated snaps for creating environments

Storage Virtualization

FC over Ethernet

Cloud Storage

Looking

Too many “needed technologies”. Not enough people+ huge storage growth Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

146

Fibre Channel over Ethernet (FCoE)

Storage Trends

Ethernet Header

Ethernet FCoE Payload

EOF

SOF

CRC

• Leverages Ethernet enhancements to carry Fibrer Channel traffic over Ethernet • Maps entire FC frame (inclusive of SOF/EOF) onto FC Frame Ethernet Ethernet FCS

Driven by INCITS T11-BB-5 with broad industry support

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://www.emulex-emc.com/downloads/Emulex-EMCLP21000_Product_Presentation.ppt 147

Storage Trends

Ethernet vs. Infiniband vs. FC ?! Ethernet's Relentless March

100000

1000

pNFS

100

FCoE?

iWarp

1

FCIP

iSCSI

10

SCSI/FCP

Time Infiniband ATM

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

FDDI

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

0.1 1990

MB/Sec

10000

Ethernet

NETAPP 148 http://www.partnersummit.com.au/2008/documents/ppts/Technology% 148 148 20Directions%20&%20Trends

Storage Trends

SFP+

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source : GlassHouse

149

FAS Deduplication in Action

Storage Trends

presentation.ppt

presentation.ppt

20 x 4K blocks = Identical blocks

Identical file 20 x 4K blocks

Data Written to Disk: With FAS dedupe: 38 blocks No FAS Dedupe: 75 blocks

presentation.ppt

job-cv.doc

http://www.netappusergroup.com/Presentations/2008November/Dedupe%20on%20Filers.ppt

Edited, 25 x 4K blocks Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Different file 10 x 4K blocks 150

Dedup for Backup environments

Storage Trends

Easy integration with existing environment

Backup and Nearline Applications

CIFS, NFS, NDMP, OpenStorage

Data Domain DD690 Appliance

Replication

Ethernet

FC = VTL

2U 2 - 6 ports RAID-6 500 GB SATA drives 5.4 to 35.2TB with shelves File system NVRAM N+1 fan Gateway to: 3PAR, EMC, Fujitsu, HDS, HP, IBM, NetApp, Nexsan, Pillar, Sun, Xiotech 151

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

151

Dedup for production storage?!

Storage Trends

• Deduplication is NOT performed in real time • ONTAP process performs deduplication – Only one deduplication process can run on a flexible volume at a time – Up to eight deduplication processes can run concurrently on a controller

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

http://www.netappusergroup.com/P resentations/2008November/Dedu pe%20on%20Filers.ppt

152

Dedupe and VMware

Storage Trends

• VMware O/S VMDKs dedupe extremely well – Maximum savings can be achieved by keeping these in the same volume – 80+% space savings possible (suitable VMDKs)

http://www.netappusergroup.com/Prese Pini Cohen’s work Copyright 2009©STKI ntations/2008November/Dedupe%20on Do not remove source or attribution from any graphic or portion of graphic %20Filers.ppt

153

Flash Solid State Drives for Tier “0” Requirements

Storage Trends

Unprecedented Application Performance Characteristics

• 30x IOPS improvement • 10x faster response time Flash Drives

• No moving parts for high reliability • 73GB & 146GB Drives Source: EMC with STKI modification Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Response Time Millisecond

• 38% less power per drive

100% Read Miss Workload

15K FC Drive Flash Drive Workload - IOPs 154

Power Savings Storage

Storage Trends

Server tells Hitachi array to power up for transaction HDDs power up Transaction takes place Server tells Hitachi array to power down HDDs power down

Power On RAID Group Power Down RAID Group Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: HDS STKI modifications

155

Blade Storage Solution : Common scenario

Storage Trends

D2D Backup Server

VMware VMware

Exchange

iSCSI Storage

SQL

Snapshots

Source: HP STKI modifications Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

156

Blade Storage Solution

Storage Trends

Shared storage, file serving & easy management in the enclosure − File/Print Serving ( Windows NAS) − Shared Application Storage (iSCSI SAN) − Integrated Data Protection (Snapshots, replication, Data Protector Express BU Software)

D2D Backup Server

AiO SB600c

VMware VMware

Replicatio n LAN/ WAN

Exchange SQL

Tape backup

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Snapshot s

Source: HP STKI modifications

157

Microsoft DPM

Storage Trends

Disk-based Recovery

Active Directory® System State

Online Snapshots (up to 512)

Up to Every 15 minutes

DPM 2007 With Integrated Disk & Tape Windows XP Windows Vista

Windows Server 2003 Windows Server 2008 file shares and directories

DPM 2007 Continuous Data Protection for Windows Application and File Servers Rapid and Reliable Recovery from disk instead of tape Advanced Technology for enterprises of all sizes Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Tape-based Archive Offline Tape

Storage Local Trends

Storage Ratios

• Number of TB (Raw) per Storage Staff Member (including backup and DRP of storage): • In Israel from 10TB per FTE (in small organization with 7TB) to 330TB per FTE in large Telco\Defence\Finance organization. • Median is 1 FTE for 100TB raw (including backup\drp) • Jump in data from 50TB per FTE to 100TB per FTE • About 25% increased ratios from last years data (for the large organizations) • Issues in storage ratios: – Net vs. Gross. What is Net (with DRP, after Raids)? – Raid 1 vs. Raid 0 vs. Raid 5 (and other Raid types) – VTL storage – Backup operation is distributed Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: STKI 159

Storage Local Trends

Storage Ratios • Storage is the area where many clients complained they are understaffed – not able to control the allocations, proactive maintenance , new technologies

• Want to have better ratios? allocate each application highend storage x4 of what was asked…. • STKI predicts – IT failures caused by Storage

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Source: STKI 160

Storage pricing data

Storage Local Trends

• Storage prices for large highend storage deal (above 3M$) is about 4K$ per raw TB including SW, HW and warranty for 3 years (most drives fiber) • Storage prices for not so large (400K$) deals are around 3K$-4K$ SW, HW and warranty for 3 years • Example of good deal (about 20K$) for midrange 500G SATA box for about 1.75K$ per TB.

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

STKI recommendation • Can I define Service\Standards for Storage, Backup, DRP ?

• STKI Round Table : Storage technologies - VTL, Dedup, Thin Provisioning , CDP at 29.06.09 Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

162

‫‪WIKIPEDIA‬‬ ‫‪STKI‬‬ ‫מתוך ויקיפדיה‪ ,‬האנציקלופדיה החופשית‬

‫חברת ‪STKI - Schwarzkopf The Knowledge Integrators‬‬ ‫היא חברת ייעוץ בתחום מערכות מידע בישראל‪ ,‬המפעילה צוות‬ ‫של אנליסטים‪ .‬לקוחות החברה כוללים בנקים‪ ,‬חברות ביטוח‪,‬‬ ‫משרדי ממשלה‪ ,‬חברות היי‪-‬טק ותעשייה‪ ,‬חברות טלקום ועוד‪.‬‬ ‫באמצע שנות ה‪ ,80-‬ד"ר ג'ימי )חיים( שוורצקוף ביצע את‬ ‫המחקר הראשון מסוגו אודות שוק טכנולוגיית המידע בישראל‪,‬‬ ‫עבור חברת ‪ .DEC‬המחקר הכה גלים בשוק הישראלי‪ ,‬ונתן‬ ‫תמונת מצב קונקרטית בשוק שהיה עד אותה עת בלתי ממופה‪.‬‬ ‫ד"ר שוורצקוף המשיך בחקר השוק ובשנת‪1992‬הקים את חברת‬ ‫‪ ,Meta Group Israel‬ששימשה כנציגות בישראל של חברת‬ ‫‪ Meta Group‬העולמית‪ .‬במסגרת החברה‪ ,‬המשיך ד"ר‬ ‫שוורצקוף לפרסם מחקרי שוק בתחום באופן קבוע‪ ,‬וכן לספק‬ ‫ייעוץ לחברות והארגונים המובילים בשוק‪ .‬בשנת ‪ 2004‬נרכשה‬ ‫‪ Meta Group‬העולמית על ידי חברת גרטנר‪ ,‬וד"ר שוורצקוף‬ ‫והאנליסטים המשיכו בפעילותם בישראל תחת השם ‪.STKI‬‬ ‫‪163‬‬

‫‪Pini Cohen’s work Copyright 2009©STKI‬‬ ‫‪Do not remove source or attribution from any graphic or portion of graphic‬‬

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

164

Announcements • This is not just another lunch!

• During lunch you will get an SMS with two questions. Please answer (before desert..). Results will be shown in Jimmy’s presentation.

Pini Cohen’s work Copyright 2009©STKI Do not remove source or attribution from any graphic or portion of graphic

Related Documents


More Documents from "W.J. Zondag"