The TCO discussion may not be the main obstacle for Mainframe revival
Posted bySven Karlsenat
2006-02-16 06:26
I've read your notes and comments on the "war" between i386-servers and mainframes with great interest.
I don't know anything about mainframes and their TCO, but I'll accept all your arguments without hesitation. And I understand your frustration: when it's just a question of logics, why won't the corporations accept the logic then?
I've been the CTO of a minor organisation for a decade. Too small for ever considering a mainframe, but even so, I've been through a scenario that may touch on the subject.
Some years ago, I had the financial opportunity of doing a major makeover of our entire system, and - beacsue I'm a nerd - I considered all possibilities. This also brought the option of mini's into consideration ... very appealing to me (nerd-factor 9), and my research showed a 3-year financial break-even horisont!
So, why did the choice fall on i386 again?
Well, to be honest, there was several points of objections from various dept's, but what made me drop the path was one single obstacle: the transition project.
I'm no coward, so I wasn't afraid to engage in a steep learning curve. Neither did the immediate difference in HW/SW cost scare me off, - the organisation was booming, so I had a lot of head-room.
But: the transition!?
1. if I was to guarantee virtually uninterrupted service delivery during the transition period, I would have to run two complete operations side by side (I'm sure it's just paranoia, but if it is, then why wouldn't any of the supplier put a big money-guarantee behind their promises of a system that would work from day one?).
2. Two complete systems running live in parallel, also means two sets of staff, again requiring the double double resources in space (offices and tech rooms), and so on.
The costs of this, plus the migration costs, would take quite a while to even out, and telling the BOD that I had a breask-even horisont of even 5 years, seemed be equal to saying that I had no grip on the finance (or perhaps they were just not typical, - perhaps BOD's in general find it ok with long term perspectives for IT investments?).
Anyhow, I decided against the mini, and got some new i386 servers with new SW. That too resulted in some migration costs, but the transition could be broken down in steps, the staff was the same and the premises could host everything during the period.
That would probably also end up as my main obstacle, if I was a CTO of a major corp., who should consider the step from i386's to mainframe.
Consequently, the proposal from a small-time CTO is, that you put a little focus on this subject when you think about how to sell mainframes to i386-CTO's ... because I'm sure there's quite a number of them who are as neardy about big boxes as I, but just aren't prepared to risk their career on a "personal whim" ;-)
Kind regards,
Sven
Migration is a bear
Posted byJohn Albeeat
2006-02-17 19:00
I'm not advocating migration to a mainframe as much as I'm advocating the mainframe is a platform too. Transition across platforms is more difficult than expansion of the same platform, but in planning for growth there might be a scenario where the move across platforms provides a more desirable result then just adding more platforms. In that case, the migration will have to happen, and redundancy will have to be there to insure continued service.
You may not be looking at such a scenario. But take a real world example of a large local bank in a closed in a somewhat isolated society, which strategically wants to deliver financial services to the world. They know they need to provide 7x24x365 availability to be successful. They have areason to consider moving to mainframes and running the fastest db they can get their hands on. And that is what a handful of China banks are doing as they break through the isolation and begin competing in the world market. I've heard rumor that they will have the largest IMS deployment IBM has ever seen. IMS is one of the oldest manframe database but it is the fastest, and those who need to process transactions at breakneck speed are using IMS.
If the rewards warrant it the risk will be managed!
Forgot your password?
New user?
I don't know anything about mainframes and their TCO, but I'll accept all your arguments without hesitation. And I understand your frustration: when it's just a question of logics, why won't the corporations accept the logic then?
I've been the CTO of a minor organisation for a decade. Too small for ever considering a mainframe, but even so, I've been through a scenario that may touch on the subject.
Some years ago, I had the financial opportunity of doing a major makeover of our entire system, and - beacsue I'm a nerd - I considered all possibilities. This also brought the option of mini's into consideration ... very appealing to me (nerd-factor 9), and my research showed a 3-year financial break-even horisont!
So, why did the choice fall on i386 again?
Well, to be honest, there was several points of objections from various dept's, but what made me drop the path was one single obstacle: the transition project.
I'm no coward, so I wasn't afraid to engage in a steep learning curve. Neither did the immediate difference in HW/SW cost scare me off, - the organisation was booming, so I had a lot of head-room.
But: the transition!?
1. if I was to guarantee virtually uninterrupted service delivery during the transition period, I would have to run two complete operations side by side (I'm sure it's just paranoia, but if it is, then why wouldn't any of the supplier put a big money-guarantee behind their promises of a system that would work from day one?).
2. Two complete systems running live in parallel, also means two sets of staff, again requiring the double double resources in space (offices and tech rooms), and so on.
The costs of this, plus the migration costs, would take quite a while to even out, and telling the BOD that I had a breask-even horisont of even 5 years, seemed be equal to saying that I had no grip on the finance (or perhaps they were just not typical, - perhaps BOD's in general find it ok with long term perspectives for IT investments?).
Anyhow, I decided against the mini, and got some new i386 servers with new SW. That too resulted in some migration costs, but the transition could be broken down in steps, the staff was the same and the premises could host everything during the period.
That would probably also end up as my main obstacle, if I was a CTO of a major corp., who should consider the step from i386's to mainframe.
Consequently, the proposal from a small-time CTO is, that you put a little focus on this subject when you think about how to sell mainframes to i386-CTO's ... because I'm sure there's quite a number of them who are as neardy about big boxes as I, but just aren't prepared to risk their career on a "personal whim" ;-)
Kind regards,
Sven
You may not be looking at such a scenario. But take a real world example of a large local bank in a closed in a somewhat isolated society, which strategically wants to deliver financial services to the world. They know they need to provide 7x24x365 availability to be successful. They have areason to consider moving to mainframes and running the fastest db they can get their hands on. And that is what a handful of China banks are doing as they break through the isolation and begin competing in the world market. I've heard rumor that they will have the largest IMS deployment IBM has ever seen. IMS is one of the oldest manframe database but it is the fastest, and those who need to process transactions at breakneck speed are using IMS.
If the rewards warrant it the risk will be managed!