Unicode vs. Non-Unicode with MI

[Former Member]

Hi all,

We are looking at doing an installation of MI. We are currently Non-Unicode on our ERP system, although a Unicode conversion project is not too far away.

According to the document: "How To Design a SAP NetWeaver Based System Landscape" (NW 7.0, ver 1, March 2008) on pg 15, it says:

"It is required that an MI system with AS ABAP has the same Unicode type as the MI back-end system: If your MI back-end system is a Unicode system, also install a Unicode MI system (that is, both AS ABAP and AS Java on Unicode); if your MI back-end system is non-Unicode, install an MI system with non-Unicode AS ABAP (that is, AS Java on Unicode and AS ABAP on non-Unicode).

Recommendation

For new installations, it is recommended to install a Unicode based system for all SAP applications and SAP NetWeaver deployments that require an AS ABAP usage type based system (AS JAVA is only available in Unicode). In future all new installations from SAP will be Unicode only."

This is clearly conflicting...We are an English-only shop, we have a Unicode conversion imminent, and SAP's direction is clearly Unicode.

Should we even consider a Non-Unicode install?

Thanks,

Troy

[Former Member]

Very helpful answers guys, this was my feeling from the beginning.

Warm regards,

Troy

[Former Member]

Hi Tro,

there is a new note avsailable in SAP system that says: you can run MI Unicode middlewarw against Non Unicode Backend - and so you are back to the first sentence: new installation - always unicode.

I am not sure about the note number, but hopefully you can find it while having a look

Regards,

Oliver

[Former Member]

Hi,

install a unicode middleware system for the following reasons:

- Soon the BE will be unicode enabled.

- The MI client is anyway unicode enabled, this means the end users can enter unicode characters anyway. After Unicode conversion of the backend your MI server would be the only non unicode component.

- having the mw as unicode does not harm, even if the BE is non unicode (apart from the bigger DB size due to the code page)

Best Regards,THomas