February 2010: NetWeaver BW 7.2 cancelled
I was pretty disappointed when the NetWeaver BW 7.2 release was cancelled, especially as it was cancelled on the day of its release. It seemed to add some nice functionality that allows large enterprises to capitalise on the latest innovations in Data Warehousing. The ability to use Semantic Partitions to easily separate and aggregate regional data to global entities is particularly interesting to anyone looking to run a Single Global Instance, because you can pull the information needed for one region into the system, whilst not affecting information needed for another region.
Anyhow I soon got over it, because George Campbell-Kelly nudged me to engage with SAP to find out what was coming next, and if we could be part of it. Some calls later I found there was a beta for NetWeaver BW 7.3 starting in July and running until November, and we could be part of it as a partner or customer. Beta programs require some investment and so I checked that we would have the buy-in from the BW and Planning teams to commit to proper testing, blogging etc. and they put some real time aside in their schedules.
NetWeaver BW 7.3 beta program
It took a bit of hacking to get the system loaded (it’s a beta after all) but after a few short hours, we had BW 7.3 loaded on a fresh sandpit. We had a problem which was that the beta comes with no Business Content so it wasn’t possibly to hook it up to source systems without moving Business Content from 7.0 systems. Which isn’t hard. What was immediately impressive though was the quality of the beta. It looks like SAP took the code line from NetWeaver 7.1 EhP2 and forked it, then pulling the code from BW 7.2 functional tier back into the NetWeaver stack, and bug fixed the resulting tree as a 7.3 release platform.
NetWeaver ABAP is a pretty mature platform by now, and it sounds like it took them 4 months to get a beta release out, which isn’t bad. In the meantime they stabilised NetWeaver to a full platform release including NetWeaver Process Integration (PI) and Composition Environment (CE). And two observations really come out:
1) It flies – in the lab
When you upgrade from BW 3.5 to 7.0 you can get massive performance increases but you have to fight for them. Moving from the 3.5 objects (ODS, transfer rule) to 7.0 objects (DSO, DTP) – which gets a performance gain of 30-40% on data loads – is a project in its own right and shouldn’t be undertaken lightly – in fact most sites do not choose to upgrade BW and convert to the new objects at the same time. But when you upgrade from 7.0 to 7.3, you do nothing and get a 30-40% additional performance increase for loads, at least this is what we found in testing. I’ll let you know later how that has worked out for us productively.
2) It’s rock solid – relatively
Now I can’t really comment on the Ramp Up, as it’s only being released today, but the Beta was rock solid, considering. That’s not to say that we didn’t find a good number of bugs and report them to SAP, but that’s why we participated in the Beta in the first place. And what’s more, anyone who has been a part of a serious NetWeaver BW implementation will know that it’s pretty normal to find program errors in a project lifecycle, that need fixing. The beta was around that level and wasn’t full of thousands of errors that hampered our every progress.
Well today, the Ramp Up for NetWeaver 7.3 commences and this will include BW, PI and CE. Bluefin’s roots are set in NetWeaver BW and so we are very heavy information users for the size of organization that we are. We have sales and opportunity information from CRM, Project Systems data for reporting on consulting performance, Finance, HR, Recruiting and Support Desk information. We believe that being an information rich organization gives us a competitive advantage. For those of you that have been following, you may know that we have Project Santa in progress, which brings proper project planning into NetWeaver BW Integrated Planning with integration back into Project Systems.
And on day 1 of the Ramp Up, we plan to be installing 7.3. I suspect this is for a number of factors.
- Being part of the beta program gives us confidence that this is a stable product
- Getting on 7.3 on gives am early competitive advantage because we know the product before anyone else – outside the lab
- We wanted to mess with the heads of the people on Project Santa and take their system down during week 1 of realization
- We have excellent test scripts (thanks Chet)
- I’m sure there’s a bit of geeky get there first going on
So I’ll be letting you all know how we get on as the week progresses. It is our Month End this week and our Finance department won’t be releasing the system back to us until Wednesday 8th December 2010 so provided our regression test goes to plan, we will be upgrading the Productive environment between 8th-10th December 2010. And if that’s the case, I’d think that we will be the first site to go live globally.
But there’s a caveat
We were part of the first global Integrated Planning project back in early 2006 and those that remember it, remember it with afflicted affection. There were many long nights, games of table football (fussball for the American audience) and we had a dashboard for “Rum & Cokes” on the wall – you get a dash for each mistake you make I’m pretty certain James O’Callaghan is still in debt here. Hugh will keep me honest. The point is we took a new product and tried to use the latest and greatest technology.
This is a lesson we’ve learnt in 7.3 and we will be upgrading our system as it is and not using the new features until they have been tested as working. This way, a Go Live won’t be impacted by what we don’t yet know about 7.3.
Hasn’t been written yet! Watch this space!