Increase release cycles for Oracle Analytics Server

Michal Zima

Description

I guess that still many customer from different reasons (for us as a central bank is the conservative approach by its nature and mitigation of potential reputation risks) are still (and will be for some time) using on-premise sw - in case of Oracle Analytics product it is OAS.

And having just on release cycle for OAS per year (and thus lagging quite behind in terms of functionality in comparison with cloud customers) is something , which we cannot appreciate that much.  When we compare to competitive products - for example Power BI Report Server (on-premise version of Power BI) is having since 2019 3 releases per year (see : https://powerbi.microsoft.com/en-us/blog/power-bi-report-server-release-timing-update/ ). So at least 2 releases per year for OAS will give all on-premises customer a feeling, that Oracle is not pushing them completely to the sidelines.

I hope, that this idea (or wish) will not be the solely expression of our institution.

Thanks and regards

Michal Zima

Use Case and Business Need

Leveraging new enhancements in Oracle Analytics, developed primarily for OAC sooner then with one year delay in on-premise OAS.

Original Idea Number: cf58d34459

Gabby Rubin-Oracle

Hi Michal,

I do not dismiss the request to have more frequent OAS cycles, but stating that an annual release creates the perception that Oracle is sidelining on-prem customers is a bit concerning to me. As some will recall, when Oracle Analytics (OBIEE at the time) was an on-prem only product, releases came every 2, 3, or even 4 years apart. In the last 2 years, we not only, for the first time in the product history, committed to annual releases for our on-prem customers but also made sure they will get all the same functionality as our cloud customers with the inclusion of the self-service and augmented analytics capabilities for all OBIEE customers for free. There is always room for improvement, but I do hope that our action demonstrates a strong commitment to our customers regardless of their deployment choice.

The reason that OAS has a longer release cycle is that the effort to create an OAS update is far greater compared to OAC. A lot of the additional effort going into specific projects like porting to Windows, testing a variety of RCU databases, and most importantly building the upgrade and LCM facilities to allow customers to upgrade from one version to the next. Unlike OAC where every deployment has an identical topology, in OAS we have 20 years of history with custom deployments when it comes to security, networking, clustering, etc. This unique effort for OAS takes months to complete and the main reason that OAS is on an annual schedule. 

We are exploring options that will allow us to accelerate, but even there not all customers will like those options. As an example, one idea is a containerized deployment of OAS, however, it would mean the introduction of some standardization into OAS deployments which might invalidate some topologies that customers use today. The equation is fairly simple, to get to faster delivery we need to reduce the number of uniquely moving parts, doing so without hurting your long-standing customers with massive custom deployments is the tricky part. 

In summary, we are goal aligned. I and my team only care about one thing, getting more value at the hands of our customers sooner. There is nothing more frustrating to a Product Manager than having a great feature but being blocked from giving it to your users. So in a sense, I'm upvoting your idea, but this is not a simple release cycle decision, OAS is an enterprise product with a high level of flexibility to allow complex customized deployments, which drives significant effort into creating deployable updates. 

Regards,

Gabby

Michal Zima

Hi Garry,

thanks very much for your very comprehensive answer - I really appreciate that. And all the argument, where you are explaining the rationales for having just one OAS release per year seems to be valid for me. It assures me (although it may seem that cloud "rules the world" and on-premise is something "old-fashioned" - but there are couple of reasons for not to move to cloud yet for us - one of them for example is the flexibility in configuration, you are mentioning), that we (as OAS on-premise customers) will (even a little bit later) get the same value as our cloud counterparts. I am also quite old OBIEE "veteran" (I became familiar with the product when his name was "Siebel Analytics"....) and I am really glad, that this product is (after some time of "trampling on the spot") again getting the right course (with all those new self-service , ML ....capabilities). Just one last point - I would be good not forget to improve (and move forward) also the "old-fashioned" OBIEE components, they deserve some innovations as well  (I noticed for example planned improvements in data modelling area - web based modeler along with semantic modelling markup language - this is something I am really curious about....).

Thanks again for your answer, which dispel (at least partially) my fears and I hope, that we "bet on the good horse" for the future :-).

Kind regards

Michal

 

Khary Dia-Kane

What about having more frequent updates for just the Self Service part?