Webinar | Make vs. Buy
When: Thursday, November 12th, 2009
Featured speaker: Brian Wood, VP Marketing, Continuous Computing
Sponsors: Continuous Computing, GoAhead, and Wind River
Boost PC performance: How more available memory can improve productivity
In-house or Outsource? Evaluating the Make vs. Buy Decision
1. A Light Reading Webinar
In-house or Outsource?
Evaluating the Make vs. Buy Decision
Thursday, November 12, 2009
Hosted by
Simon Stanley
Analyst at Large, Heavy Reading
Principal Consultant, Earlswood Marketing
2. Speakers
• Brian Wood
VP of Marketing
Continuous Computing
• Bill Yaman
VP of Marketing
GoAhead Software
• Jessica Schieve
Director of Networking Solutions
Wind River
3. Light Reading Make vs Buy Survey
• Conducted during October 2009
• Based on >100 individuals at key NEPs
• Broad geographical coverage
• 44% US and Canada
• 25% Europe
• 15% Asia Pacific including China
• Big and small NEPs
• 38% > $1 Billion
• 27% < $50 Million
• Range of roles
• 42% development and engineering management
• 37% sales and marketing
5. Telecom Market Trends
•Operators are challenged
•Mobile traffic 2X every 9 months
•New capacity growth only 20%/yr
•Operators focused on increasing
revenue & reducing OpEx / CapEx
•NEP focus: key differentiation
•Application development
•Managed services
•Customer support
•Shift to standardized platforms
•ATCA, AMC, MicroTCA
•CG Linux OS
•SA Forum middleware
6. Typical Product Development Cycle
45%
40%
35%
30%
25%
Access/CPE Systems
20% Edge Systems
Core Systems
15%
10%
5%
0%
6 months or 6 months - 12 months - 18 months - 24 months - Longer than
less 12 months 18 months 24 months 30 months 30 months
7. Product Cycles and
Requests for New Features
• Product cycles are getting shorter
• Historically time to 1st trial has been
2-3 years 45%
• This is no longer a viable model 40%
• Majority from survey (>60%) 35%
30%
• < 18 months from inception 25%
Access/CPE Systems
to first trial for core systems 20% Edge Systems
Core Systems
15%
• < 12 months for Access & CPE 10%
systems 5%
0%
6 months or 6 months - 12 months - 18 months - 24 months - Longer than
less 12 months 18 months 24 months 30 months 30 months
• Frequent requests for new features
• Quarterly
• 68% Access & CPE
• 57% Edge and Core
8. Platform Evolution
Less Costly but
High Cost Best ROI
Difficult
Application Application
APIs
Vertically Mgmt Protocol
Integrated System SW
Approach OS MW ES
HW HW HW HA Platform
Proprietary platforms Horizontally-integrated Standards-based system
In-house R&D + Integration “Plug-n-Play” fallacy Enables focus on value-add
Shift From Vertical to Horizontal Integration
9. What Are NEPs Most Likely to Purchase?
Best In-Class
Building Blocks
Pre-integrated
Hardware
Lowest Price
Building Blocks
Application Ready
System
Complete System
0% 10% 20% 30% 40% 50% 60% 70%
Likely or Highly Likely
10. What Drives NEPs’ Decision to Buy a
SW / HW Platform vs Develop?
60%
50%
40%
Time-to-market
Cost
30%
Number of products available
Functionality
20% Carrier-grade capability
Standards
10%
0%
Not Important 2 3 4 Critical 5
1
11. Platform Evolution
High Cost, Better TTM, but Best TTM,
Slow TTM Costly & Difficult Best ROI
Application Application
APIs
Vertically Mgmt Protocol
Integrated System SW
Approach OS MW ES
HW HW HW HA Platform
Proprietary platforms Horizontally-integrated Standards-based system
In-house R&D + Integration “Plug-n-Play” fallacy Enables focus on value-add
3 Years 2 Years 1 Year
Shift From Vertical to Horizontal Integration
12. Time to Market
Proprietary Platform
Application System
Platform Development Field Trials
Development Integration
Time to revenue (33-48 months)
ATCA Platform
Platform Application System
Field Trials
Development Development Integration
Time to revenue (23-39 months)
Application Ready ATCA Platform
Plat Application System
Dev Integration
Field Trials
Development
Time to revenue (18-29 months)
6 12 18 24 30 36 42 48
Months
Source: Earlswood Marketing
13. Achieving Measurable Results
Development savings: 70 person years
($10M)
Reduced time-to-market from 36 to 14
months
Went from “zero” to commercial
deployment in 9 months
14. Do NEPs Use Third Party Software?
Value add &
• > 50% outsource at least 20% differentiation
of system software Applications Vendor D
• 10% outsource 80-100% of
system software
Integrated COTS Vendor C
Middleware
• What are NEPs Outsourcing? Vendor B
• Most NEPs outsourcing operating Standard O/S
system
Vendor A
• Many outsourcing protocol stacks Standard Hardware
and middleware
• Some (~15%) outsourcing COTS-Based Platform
applications and network
management
15. Example NEP Solution
Voice &
Connectivity Session Service
Applications Control, Management &
Intelligent IP Charging
Edge
Component-based MW to match the product Domains
Majority is open source & COTS
Middleware Key functionality is in-house
In-house integration software
Standards always when applicable
Operating
Linux (mainstream), other commercial, in-house
Systems
Hardware Focused ATCA IT Hardware
Hardware
Source: Nokia Siemens Networks
16. How Often Do NEPs Upgrade Software?
45%
40%
35%
30%
25%
Access/CPE Systems
Edge Systems
20%
Core Systems
15%
10%
5%
0%
Quarterly Twice a year Yearly Longer
17. Audience Poll
Is anything stopping you from meeting your
software upgrade goals?
• No, we’re meeting our upgrade goals
• Yes, in-service upgrade restrictions
• Yes, limited development resources
• Yes, time to test upgrades
• Yes, other issues
18. How Important is Standards
Compliance?
3%
Very important
21% Somewhat important
Not important at all
76%
• Software Standards Deliver
• Platform Independence
• Application Portability
19. Carrier-Grade Attributes Required
Throughout the Entire System
3%
Very important
23% Somewhat important
Not important at all
Carrier-Grade Applications
High Availability Middleware
74% Carrier-Grade Linux
or RTOS (VxWorks)
COTS-based Hardware
• Highly Available • Scalable
• Standards Based • High Performance
• Serviceable • Secure
20. ATCA Usage
• Are NEPs Using ATCA?
• >60% use ATCA for some systems
• ~10% use ATCA for 80% of systems
• Is ATCA COTS or In-house
• ~40% is all in-house design
• ~10% are 80-100% sourced
from ecosystem
• Market Data from Recent Heavy Reading Report
• ATCA-based systems will be almost $800M in 2009
• COTS will be $423M in 2009
• Roughly 50/50
21. What Is Important to NEPs When
Considering ATCA?
Critical or Important
Open standards
Best-in-class components
Total system solution
CapEx
Feature differentiation
OpEx
Ecosystem competition
Vendor choice
0% 10% 20% 30% 40% 50% 60% 70% 80%
• Open standards • No vendor lock-in
• Max ecosystem innovation
• Leverage Moore’s Law
• Best-in-class components • Multi-core SW optimization
• Focus on app software dev
• Total system solution • Use common infrastructure
• CapEx • Pay-as-you-go growth
• Skip lengthy & costly R&D
22. Conclusions
1. Time to market is more critical than ever
• 12-18 month product development cycles
• Rapid introduction of new technologies
• Shorter time to revenue
2. Standard platforms enable faster time to market &
reduced cost of ownership
• Best-in-class components
• Application ready platforms
3. Off-the-shelf software & pre-integration further
accelerate system development
• Operating system, protocol software & middleware for many
• Applications and network management for a few, but growing
4. ATCA used for some systems by more than 60%
• 55% source some components from ecosystem
24. Thank you for joining us today!
For an archived version of this Webinar,
please go to the Light Reading Webinars page:
http://www.lightreading.com/webinars.asp