Unicode-UTF8-Support for MaxDB?

[Former Member]

Hello,

according to general information MaxDB just supports UTF16. If one is

facing the decision to migrate an Oracle 9.2 Non-Unicode-DB to either a Unicode-UTF16-MaxDB or a Unicode-UTF8-Oracle-DB, the answer probably would be to go for a Unicode-UTF8-Oracle-DB because the SAN costs used should be app. half of the MaxDB-SAN-costs.

Is UTF8-Support soon ready for MaxDB?

Thanks for your reply.

br

Chris

[roland_mallmann · Accepted solution]

Hi Chris,

MaxDB currently uses UCS-2. This decision was made some years ago and was influenced partly by 1) Microsoft's decision to go with this and 2) to handle the asian market well (UCS-2 asian characters are stored more compact than when using UTF-8).

There are ofcourse some downsides (size-wise) when comparing UTF-8 and UCS-2 in the euro-us market, but it would never come down to exact a ratio of 1:2 (UTF-8:UCS-2). I cannot give you any numbers though.

On the plus side, MaxDB can currently support a mixture of ASCII and Unicode, if needed.

Regarding UTF-8 support; it's still on the agenda, but long-term.

Kind regards,

Roland

[Former Member · Accepted solution]

Hi Roland,

thanks for your answer.

I found OSS 914966 saying that the occupied space is up to 50% more with USC-2 in comparison with UTF8. This means that with a 2 TB-DB under UTF8 it would take 3 TB under USC-2. If in addition there are different test systems that are copies of the production system, the additional SAN-storage requirements for USC-2 is even bigger.

If a company looks (and it normally does) at the lowest TCO (total costs of ownership) then the bet would be for Oracle in this example.

Wouldn't it be helpful to soon support UTF-8 on MaxDB so that a big number of customers would change to MaxDB (that must be in also SAP's interest)? If it is not the case soon, what is the reason?

You say that MaxDB can currently support a mixture of ASCII and Unicode, if needed. Where is this procedure described and is it officially supported by SAP?

Thanks for your feedback.

Kind regards,

Chris

[roland_mallmann · Accepted solution]

Hi Chris,

yeah, that note needs a small change (UTF-16 changed into USC-2).

Indeed, a 2TB DB under UTF-8 needs about 3TB under USC-2. So yes, one would need more space for MaxDB, but the TCO is not only dependent on the cost of the SAN. Easy admin, being reorganisation free are only a few of the examples which can drastically lower the TCO of a system. That added to the low cost of the MaxDB software itself <i><b>can</b></i> make a difference.

Because of other topics, which I cannot go into and currently have a higher priority, the UTF-8 support is planned long-term.

I'm not sure if this is allowed, but this (german) article is quite an interesting read:

<a href="http://www.computerwoche.de/produkte_technik/storage/581295/">http://www.computerwoche.de/produkte_technik/storage/581295/</a>

---

With the Unicode / ASCII mixture I just meant that for example a Unicode WebAS contains ASCII data aswell. Another example: in MDM there's an explicit distinction between the ASCII and UNICODE data. Our interfaces support all three (ASCII, UTF-8 and UCS-2).

Regards,

Roland

[Former Member]

Hello Roland,

thanks for your answer. I'm afraid the many customers will stick with Oracle and not change to MaxDB due to the mentioned additional storage requirements that are not to be neglected if you look at a SAP system landscape (portal, xi, erp , bw, Sol) with 20 TB worth of SAN.

This way SAP misses an opportunity to draw Oracle SAP customers to the MaxDB side.

Best regards

Chris