LaoTsao's Weblog (老曹的網路記)

Solaris for AMD and intel chipsets

Solaris for AMD/Intel

There is renew effort to push Solarsi on AMD/Intel chipset.

The ma jor issue is, how to get ISV to port their apps to Solaris for AMD/Intel?

One is like building a field of dream, you just build it and people will come, so when ISV see the advantage they will port their apps to solaris/AMD/Intel

One is for SUN to help the porting effort

1. providing money
2. providing people resource

IMHO, SUN will need to spend about $M for 1000 Top ISV This is just half of the money SUn get from MSFT

Posted at 10:23AM Jul 03, 2004 by hstsao in General  | 

helping kids in rural china

Helpig kids in rural china

for many years , my wife Dr. Sou-Tung Chiu-Tsao and I use Education and Science Society, INC to funnel some small amount of money to help

• adopt a rural library project $600
• aid for middle school student $100
• aid forhigh school student $180

we have contributed to many library and 20+ kid in school, we will continue to do this as long as we can affort it.

please visit Education and Science Socity and participate

thx

Posted at 07:54AM Jun 27, 2004 by hstsao in General  | 

open source benefits

Open Source

there are renew talk about openesource solaris now, this is a good thing. From the experience fron BSD and now GNU/LINUX , apache project, we see many benefit from opensource project. There are many more project at SUN, IMHO, can be opensource:

• How to write x86 Drivers
• How to write a module for particular HW in Sun Control Station
• How to write module in N1Grid SPS SW
• How to write Cluster agent in Sun Cluster (SunPlex agent Builder fill the major gap already)
• How to write module in SunMC for different OS and HW
• ......

why open source

• Open source project force project team to explain thing to much large audience
• more user can contribute to project that project team does not have the resource to do
• more user feel that they have more control on the product that they relay on for their business
• win win situation
• one can almost use the situation "free world" vs "not free world" arguement for open source

Posted at 06:46AM Jun 27, 2004 by hstsao in General  | 

what if answer

Answer

as SE work for SUN 10+ years and a customer since 1985, I have some whatif questions that wish someone can comment:-)

1. * What if SUN has Sparc CPU that clock at 4Ghz Now

   SUN will be still the king of the WS and server

2. * What if SUN open source Solaris 5 years ago??

   linux will never happen

3. * What if SDS has a GUI and be part of Solaris eight years ago

   veritas will still be a small company

4. * What if SUN keep NetDynamic as AS?

   BEA?

5. * What if SUN sale 1U or 2U server based on x86 chipset 4 years ago

   Dell will be just a laptop and WS company

Posted at 06:11AM Jun 17, 2004 by hstsao in General  | 

What IF

What IF

As SE work for SUN 10+ years and a customer since 1985, I have some whatif questions that wish someone can comment:-)

• What if SUN has Sparc CPU that clock at 4Ghz Now
• What if SUN open source Solaris 5 years ago??
• What if SDS has a GUI and be part of Solaris eight years ago
• What if SUN keep NetDynamic as AS?
• What if SUN sale 1U or 2U server based on x86 chipset 4 years ago

my answer on the following blog

Posted at 04:08PM Jun 16, 2004 by hstsao in General  | 

Service Revenue Mirage

Service Revenue Mirage

Many SUN sale go through our partner, a deal involves

1. Sun HW/SW
2. HW/SW support Service contract
3. third party HW/SW
4. installation service
5. professional Service

The Service Revenue that one talks about are Item 4 and Item 5.

• Item 1 and 2 will go to SUN.
• Item 3 will not go to SUN.
• Item 4 usually will not go to SUN
• Item 5 sometime may go to SUN.

Many partner want to use their SE for Installation and professional service and they will collect those service revenues themself.

In case SUN get the professional service deal, if we contract out the work to our partner, then we will only collect a small piece of the action.

To collect big service revenue, SUN will need to do the whole thing ourself and do many repeatable business in a short and efficient way.

Net service revenue is not easy thing to collect. I have a friend used to work at XXX, XXX has a big Consulting org, he said even at XXX they are worry about the revenue come form PS. For new technology, you will need to hire outside consultants, and they cost big buck. To use your own people, you will need to train them continuiously, this cost money also.

Posted at 03:41PM Jun 16, 2004 by hstsao in General  | 

pe-sale vs post -sale function of a SE

There are many discussion about the pre-sale vs post-sale role of a SE. As a 10+ years pre-sale SE at SUN, I have a few observations

pre-sale vs post-sale

10 years ago, when I was hire as a pre-sale SE, my manager told me that, even through you are pre-sale SE but 80 % of your work will be post-sale.

• We help customers to setup jumpstart server to JS x00+ machines
• We give Solaris training for SA that had Other UNIX OS experiences
• We and SSE setup Autoclient for 400+ WS connecting to 2 S6000 Server
• we burn many mid night oil with the back-end eng for autoclient problem
• We have full access to the datacenter and root access to the servers
• We are always at customer site to answer questions and solve problems and architect for new project

As a pre-sale SE, we donot charge our service , but IMHO our pay out is in the form of the next big order

As I moved to a different city and cover different market and different customer,

• We have an office at customer site
• We regularly give talks on new subject
• We answer questions and help solve problems
• We help customer architect the new system for new project
• Intsallation are most done by our SSE now
• We are always involved in the escalation process

We have build a very strong relatioship with our customers, we work together to architect the system and quote the HW/SW for their need and we are always soft sale SUN's HW/SW.

Since Sun's product are more comlex now, we will need to pick and choose the area that we learn and just a setp ahead of the customers' need.

This approach seems to work but not very scalable, there just no time to do this for every customer that I am suppose to cover.

What can be done better?

• Better document what we propose for what application and why
• Better document what we develop for customer
• Better environment for us to tryout new ideas

Posted at 11:20AM Jun 11, 2004 by hstsao in General  |