Server to MIPS ROI Analysis and Forecasts, 2009-2014
WinterGreen Research announces that it
has a new model driven analysis on server to MIPS ROI. The 2009 analysis has 11 pages of
metrics driven off transaction size and type assumptions entered by the user. There are 460
proprietary workload size and type metrics associated with the various servers and 2000
different IT labor categories and geographies represented in this core module.
The server to MIPS ROI is a core module for 15 other of the models contained in the Enterprise
5 set of IT data center models. Worldwide markets are poised to achieve significant growth as
application servers continue to benefit from the enterprise need to build out e-commerce sites
that support a brand. The price for the core module is $20,500 and for each separate
accompanying model $9,980.
The price for the entire set is models is $68,500. ROI /TCO are essential aspects of decision
making in the data center. This set of models prepared by independent analysts provides an
essential guide to understanding all the variable costs that are associated with implementing
automated process.
Distributed servers proliferate in the data center. As VMWare and other virtualization
technologies take hold, there is a need to analyze the metrics associated with application
platform decision making. This set of models permits IT managers and CFO organizations
participate in the understanding of the costs in a realistic manner, taking into consideration all
the metrics that might be associated with a particular situation.
The models can be used for chargeback to the line of business, for predictive analytics relating
to the impact of Intel component and IBM Systems z technology roadmaps. e used to build a
Web page and shopping cart for e-commerce. Application servers offer e-mail, chat, and
phone for personalized web reach of product displays and shopping carts. Retailers can use
application servers to manage e-commerce and service interactions over the Web using tools
that support one continuous relationship. This expansion of application server capability
vastly improves the reach of e-commerce sites.
The System z models and the distributed server models are provided in separate lists along
with the benchmarking, workload metrics relative to size and type, and pricing that are
appropriate for each particular model offered to the market.
Labor costs from around the world are provided, showing the cost by category and job
description from 100 different countries. The models have a currency converter so the ROI can
be looked at for different countries and geographies.
Server to MIPS ROI provides insight into shared workload metrics which are calculated to be
$8 per MIP per day without looking at the shared workload impact and go down to at $2.30
per MIP per day when the shared workload value is calculated. Distributed servers operate at
$15 per GHz per day typically. The models provide a way for IT to use their own transaction
volume numbers along with choosing the particular servers and System z that might be in their
own configuration to develop a view of data center costs.
The system to MIPS ROI model is used to look at comparisons of different situations under
different circumstances. The analyses are based on proprietary collections of real data from
real clients. One example of the analysis of $ per MIP per day for z GHz vs. distributed server
GHz is as follows:
One application:
System z Server
Application cost per year
$ 85,306 cost $402,980 cost
3.2 GHz 40.5 GHz
Application cost per GHz
$ 23,707/GHz $8,592/GHz
Here you see that the cost per GHz is significantly lower for one application as intuition would
dictate. However, the System z GHz runs a lot more efficiently, so ultimately the System z is
less expensive. System z runs the application in 3.2 GHz while it takes 40.5 GHz to run the
same application on a particular distributed server. Power and cooling savings are a big part
of the System z efficiencies.
Ratio:
$23,707 System z 12.65*$8,592=$124,584
IT Labor Categories /Geographies
IT Labor Geographies
U.S.
Canada
Mexico
Switzerland
Denmark
Finland
Germany
Spain
Italy
Norway
Sweden
Ireland
U.K.
Russia
Netherlands
Belgium
Egypt
United Arab Emrites
Japan
China
Korea
Taiwan
Indonesia
Australia
Brazil
Chile
Venezuela
Argentina
South Africa
France
IT Labor Categories
IT Manager
VP - Chief Information Officer (CIO)
VP - Chief Financial Officer (CFO)
Sr. Systems Engineer
Program Manager
Sr. Software Engineer
VP - Information Services
VP - Technical Services
Director - I.T. Planning
Director - Production/Data Center
Director – Systems & Programming
Manager - Computer Operations
Manager - Security and Workstations
Manager - Network Services
Manager - Operating Systems Production
Manager - Production Services
Manager - Production Support
Manager - Systems and Programming
Manager - Technical Services
Manager - Training and Documentation
Manager - Internet Systems
Manager - Voice and Data Communications
Capacity Planning Supervisor
Change Control Supervisor
Computer Operations - Shift Manager
Computer Operations - Shift Supervisor
Manager - Database
Manager – Data Communications
Data Entry Supervisor
Manager – Data Warehouse
Manager – Customer Service
Manager - Office Automation Applications
Production Control Specialist
Production Services Supervisor
Project Manager - Applications
Project Manager - Distributed Systems
Project Manager - Network Technical Services
Project Manager - Systems
Supervisor - Hardware Installations
Supervisor - Microcomputer Support
Supervisor - Network Services
Voice/Wireless Communications Manager
Webmaster
Change Control Analyst
Computer Operator
Data Center Facility Administrator
Data Entry Clerk
Data Security Administrator
Database Specialist
Disaster Recovery Coordinator
e-Commerce Specialist
Forms and Graphics Designer
Internet Developer
IT Planning Analyst
LAN Applications Support Analyst
Librarian
Network Control Analyst
Network Services Administrator
Network Technician
Object Visual Programmer
Operations Analyst
Personal Computer Specialist
Production Control Analyst
Programmer/Analyst
Senior Network Specialist
Software Engineer
Systems Analyst
Systems Programmer
Systems Support Specialist
Technical Services Specialist
Technical Specialist
Voice/Wireless Communications Coordinator
Web Analyst
Companies Offering Servers and Mainframes
Hewlett Packard
Dell
Oracle / Sun
IBM
Fujitsu / Siemens
Report Methodology
This ROI model is part of a series that provides IT systems financial planners access to information that supports analysis of all the numbers that impact management of a large and complex data center. The methodology used in the models relates to having a sophisticated analytical technique for understanding the impact of workload on processor consumption and cost.
Both transaction size and type have an effect on the cost and utilization of processing power. WinterGreen Research has looked at the metrics and independent research to develop assumptions that reflect the actual anticipated usage and cost of systems. Comparative analyses reflect the input of these values into models. A further type of workload relates to web services invocations. Here the assumptions and values used in creating cost analyses models are based on measure of real workload and measurements made with great precision.
With respect to the value of workload offload, WinterGreen Research has worked closely with a variety of vendors, including Progress Software Data Direct and customers are referred to those models that leverage the Willhoit constant to provide accurate assessment of relative workload costs on System z vs. the zIIP and zAAP offload engines.
The models make extensive reference to existing published independent benchmarks. While these benchmarks are available as a reference in the models, the calculations in the models utilize WinterGreen Research benchmark adjustment variables to provide workload settings that account for differences in server and System z capabilities.
There is a systematic use of parallel analyses of utilization and cost for the server and the System z pages. Each assumption on a server page has a parallel assumption on a System z page. Each calculation on a server page has a parallel calculation on a System z page. There are many different views of the information to make it useful for it analysts to understand the information being presented.
Some of the most popular views are $ per MIP per day and $ per CPU cycle (in GHz) per day. Another view of the data is a comparison of workload server GHz for an application as compared to system z GHz consumed for the same application. The analysis is then extended to provide a view of $ per GHz for a particular server, and $ per GHz for a particular system z model for a particular application and workload.
There are analyses of $ per MIP per day and $ per CPU cycles per day. There are calibration calculations that expose the values of the workload assumptions underlying all the calculations. There is a separate calculation section on every page that lists the calculation performed in order, the values and intermediate values of the calculation, and the words that describe the calculation in order that the calculations are performed. Many other user friendly aspects of the models contribute to a rigorous it financial analysis.
The variables and assumptions provided in the ROI models are based on extensive experience in providing ROI to large enterprise data centers. The ROI models have lists of servers from different manufacturers, Systems z models from IBM, and labor costs by category around the world. This information has been developed from WinterGreen research proprietary data bases constructed as a result of preparing market research studies that address the software and hardware businesses.
About the Company
WinterGreen Research, has a unique research strategy that relates to identifying market
trends through reading and interviewing opinion leaders. By reading the electronic equivalent
of 40 feet of paper, WinterGreen Research senior analysts can learn a lot more about markets, a
lot faster than can be learned through expensive surveys and focus groups. Thinking about
market trends is a high priority at WinterGreen Research. As with all research, the value
proposition for competitive analysis comes from intellectual input. With a strong international
market presence and a proven track record of reliable analysis and accurate market forecasts,
Wintergreen Research offers comprehensive ROI
As product cycles have accelerated to six month time frames, forecasting accurately
becomes an essential aspect of participation in technology markets. That is what WinterGreen
Research is all about: looking at the input from reference accounts to provide accurate
competitive analysis.
WinterGreen Research, founded in 1985, provides strategic market assessments in
telecommunications, communications equipment, health care, Internet and advanced computer
technology. Industry reports focus on opportunities that expand existing markets or develop
major new markets. The reports assess new product and service positioning strategies, new and
evolving technologies, and technological impact on products, services, and markets. Market
shares are looked at in the context of impact on ROI. Leading market participants are profiled.
The principals of WinterGreen Research have been involved in analysis and forecasting of
international business opportunities in telecommunications and advanced computer technology
markets for over 30 years.
About the Principal Analysts
Ellen T. Curtiss, Technical Director, co-founder of WinterGreen Research, conducts strategic and
market assessments in technology-based industries. Previously she was a member of the staff
of Arthur D. Little, Inc., for 23 years, most recently as Vice President of Arthur D. Little Decision
Resources, specializing in strategic planning and market development services. She is a graduate
of Boston University and the Program for Management Development at Harvard Graduate
School of Business Administration. She is the author of recent studies on worldwide
telecommunications markets, the top ten internet equipment companies, the top ten contract
manufacturing companies, and the Top Ten Telecommunications market analysis and forecasts.
Susan Eustis, President, co-founder of wintergreen research, has done research in
communications and computer markets and applications. She holds several patents in microcomputing
and parallel processing. She has the original patents in electronic voting machines.
She has new patent applications in format varying, mulit-processing, and electronic voting. She
is the author of recent studies of the services oriented architecture SOA, Mid IR sensors,
content management, mid size business middleware, worldwide energy markets, solar utility
markets, solar technology markets, thin film battery markets, and webcam markets. She also
writes about regional bell operating companies' marketing strategies, internet equipment,
biometrics, a study of internet equipment, worldwide telecommunications equipment, top ten
telecommunications, digital loop carrier, web hosting, web services, nanotechnology, and
application integration markets. Ms. Eustis is a graduate of Barnard College.