The underlying data is 95% the same across both products. The main deliverable will be QlikView but QlikSense to our analysts (without tech knowledge).
We operate on VMware and I was proposing to setup:
- DEV/UAT server (QlikView mostly)
- PROD QlikView
- PROD Qlik Sense
Does anyone see any holes in this?.... I know we will have to remodel the presentation layer for Sense but am prepared to wear this to get it into the business in the short-term. Longer term we will migrate away all to Sense.
Note there are some limitations on sharing QVDs.
- Sense can read QVDs created by View. There is some overhead here, as the QVD is first converted internally to Sense format.
- View cannot read QVDs created by Sense.
If you are reusing QVDs and want to have a single Extract & Transform layer, here's a recommended approach.
1. Create the transformed QVDs uisng View.
2. Using Sense, LOAD * and STORE all qvds to create a Sense QVD copy. This is only to avoid repeating the conversion overhead for multiple reuse.
These sub-notes are a little bit confusing, let me explain a little.
- Firstly, this is only in reference to client browser requirements. So the "virtual environment" that is being referenced in subnote 3 is the virtualisation of the client PC, not the Qlik Sense server. An example scenario would be where the user was working in a citrix client session that was hosted on a virtual citrix server.
- Note that in subnote 2 it says that Qlik Sense shall work out of the box with default browser settings. Subnote 3 must be interpreted in relation to this statement. So if the user was using a client PC that was virtualised, then the default settings of Firefox might cause an issue, as hardware acceleration is defaulted on for Firefox.
- If a user is using a virtualised client PC, AND is using Firefox as a browser, then hardware acceleration can be turned off for Firefox as per this link Disable Firefox's Hardware Acceleration to Fix Slowness (and More)
Please let me know if you have any further concerns.