Mainframe, MVS and zOS Discussion

Hello,

In my project we have DB2. I am wondering if a DB2 running on a mainframe and Oracle running on Sun OS or may be any other operating system can give the same performance? I can understand that this is not a very good comparison aas I am comparing 2 different Databases and operating system but my intention is to understand that can we acheive the same speed and reliability in a different database running different operating system and vice versa??

Mainframe compnaies are not the only comapnies who are dealing with a large volume data there are open system companies also and they donot use mainframe, so how do they satisfy customer needs in terms of spped and reliability?

Should I provide more information on this? Please suggest.

You are asking a very broad question -- Googling database performance returns about 502 million hits. Frequently non-mainframe platforms improve performance by using multiple servers for the data. There are ways -- admittedly not perfect -- to compare database performance on different platforms (Google TPC for example) by using benchmarks. Note that you need to consider reliability, availability, scalability individually when looking at database systems.

Thanks for your reply.

But mainframes are still around and the data is increasing on them. Then how a new database system might be considered in replacement fo the databases on mainframes... google searches do not really this. But I think, there must have been some study done on this...

Have you tried googling along the lines of "migrating from DB2 on mainframe to PC or server based database" ?

Some sites have replaced mainframes with other platforms (usually due to perceived cost -- mainframe costs are large and very visible whereas other platform costs tend to be spread out and not very visible even though the mainframe costs and other platform costs tend to be equivalent). Such projects usually are estimated by management to take 3 years and actually take 7 to 10 years (and some of them get cancelled as the true scope of the project becomes apparent). There is a wide variety of machine sizes needed -- a large bank needs a lot more computer power, for example, than a single retail store. The larger the use of computers, the more likely the site will be using mainframes and the less likely the site will move to another platform. Small to mid-size users find it easier to move to another platform (changing their data base along the way) as the conversion costs are smaller.

In such decisions, the data base tends to be a small piece of the overall picture. Generally the decision will be made to change platforms, then the decision is made as to which platform to move to, and only then will the data base be selected.

Thanks guys, this helps me to understand it more clearly.