In a world wide press conference - open to the public - the recording is here - SAP's Supervisory Board Chairman Hasso Plattner, it's CTO Vishal Sikka (blog here) and its CMO Jonathan Becher - unveiled the SAP Hana Enterprise Cloud.
![]()
A Historic Event
This is the first time I have seen senior SAP executives talk about key cloud benefits on a SAP homegrown product - albeit not mentioning the cloud.
- Instant availability
Multiple times the long procurement (8 weeks was mentioned) for Hana hardware came up. At one point Plattner said - that this delay disrupts the Hana projects. And with SAP making Hana available in the cloud - you can be using Hana much faster, you are now only constrained by data upload time.
- Forget sizingAs with any new server technology - sizing appropriately is always a challenge. Plattner specifically mentioned that SAP with Hana Cloud now can re-size the RAM used by a subscriber. I assume SAP has seen some wide variations of the up to 5x compression factor, depending on the data to be compressed
- Automatic UpdatesEqually Plattner hit home on the faster release cycle capability of a cloud system over an on premise system.
- Complexity Masking / ReductionThe new Cloud Frame administration software will mask a lot of the complexity of running your own hardware in house - and customers will not need to come up to speed with the low level intricacies anymore
- Try before BuyA few times the trial aspect came up -just upload your data and see what Hana can do - if you like it - keep Hana around and build more on it - if not disengage again. This is of course hampered by the SAP licensing model at this point - but see more on that below.
- Better performance & elasticityPerformance on the larger boxes that SAP has procured is superior to smaller boxes enterprise may purchase. And the usage can grow and be reduced as needed.
Why all the talk on latency?
Since a long time Hasso Plattner has SAP engineers work under the 1 second performance deadline (remember that myself...), that any enterprise application should return control back to the users under one second. For a long time Hasso (correcly) maintained that due to network latencies, the 1 second performance mandate could not be achieved if servers would not be on premise solution. But now both executives claimed that SAP has managed to keep latency to 120 ms round trip to any data center of SAP - leaving close to 800 ms to application compute time, enough time for the Hana architecture to return meaningful results.PaaS or IaaS?
Plattner also mentioned that Hana ships with a number of application libraries - not only functional to populate data and analytic libraries, but also vertical KPI libraries for e.g. healthcare and oil & gas industries. This was a clear stab to SAS CEO Jim Goodnight claiming that it's not enough to just have a database that he recently made at the SAS user conference.
But overall SAP does not offer enough packaged development tools for 3rd party developers to start building their applications on Hana on the cloud today. My guess is we need to wait for Netweaver (or Hana) Cloud products to be bundled and made available in the cloud - maybe even only to Sapphire next week. And despite Plattner claimed that you can use any programming language and application server on top of Hana, it would be good to see that as supported offering and being uptaken by enterprise software vendors beyond SAP.
Until then - I see Hana Cloud clearly (still) on the IaaS side.
 |
| Screen Shot of Hana Cloud Frame |
True next generation applications - we are (still) waiting
This remains my sore spot with SAP - again all examples mentioned were traditional performance problems (manufacturing runs, distribution optimization, pricing yield management etc) - nothing though provoking, innovative on how an in-memory enabled enterprise should operate in the 21st century. You should notice that 95% of the event was about technology, and applications were marginalized.
But I am still hoping for Sapphire bring up more though leadership in the area of in memory best practices. (see my Sapphire wish list here).
But I am still hoping for Sapphire bring up more though leadership in the area of in memory best practices. (see my Sapphire wish list here).
Still Terra Incognita - or better nubes incognitae for SAP*
While it was great to see SAP embrace a good number of the cloud benefits as mentioned above, their remain a number of key cloud capabilities and practices that SAP has been mum on. Again this may change already next week, but here are the key misses:
- While the deployment is now elastic - the licensing isn't. SAP requires customers to bring their own license - and has no way to scale up or scale down the licensing usage of Hana, based on the customer's needs. This is part of the overall challenges of a license vendor moving from license revenue to subscription revenue - not an easy nut to crack. So while SAP gets the disruption on the technology side - it needs to step up to make the business side, too.
- SAP rightfully made the elasticity claim - but with booking capacity on a monthly level - with industry best practices being on an hourly level with Amazon's AWS - the usage pricing model needs to be equally adjusted.
- As mentioned, Hana Cloud is more IaaS than PaaS and far from SaaS. SAP ultimately needs to be a SaaS vendor and equally rethink the licensing and release policy of its business applications. For SAP to lead a cloud initiative from the bottom up and not the top down - remains something that keeps puzzling. Should SAP even talk about technology - or instead talk about great enterprise applications?
- It remains unclear who will run the data centers - SAP themselves or SAP partners. Since cloud requires enterprises to do a significant leap of faith - SAP will need to provide more information on the deployment side. If SAP chooses the partners and avoids the capex hit - it will slow down speed to market and reduce SAP's control over the technology adoption - and ultimately its success.
- Cloud integration is one of the recent buzzwords in the market, but also a real world problem for companies moving to the cloud. Nothing addressed by SAP at the event, the challenge seems to be that Netweaver (or Hana) Cloud platform are not ready yet (again we will see next week at Sapphire).
The good news for SAP is, that assuming Hana Enterprise Cloud will be a success, that all of the above will need to be addressed successfully, so Hana Enterprise Cloud revenue keeps thriving. It's always good to have revenue goals helping to overcome roadblocks of other nature.
MyPOV
With RAM storage remaining the most expensive piece of Hana hardware, using the cloud to make Hana applications more elastic, is a smart move by SAP. If all enterprise data though will move to this expensive cloud medium - will remain to be seen. The Oracle flash based plans for Oracle 12c are more affordable in the next years, but I hope nothing limits SAP to deploy Hana equally on flash memory.
Moreover, being successful with a IaaS product in the database area is a challenge, a challenge so big, even better marketing and sales driven organizations (aka salesforce) have not made a killing with their IaaS products (database.com). And SAP needs to pick up a lot of speed - as with all things in the cloud - Amazon AWS is the benchmark, redshift is the hurdle - and with Infor endorsing it - redshift is noticeably ahead of Hana at this point.
Moreover, being successful with a IaaS product in the database area is a challenge, a challenge so big, even better marketing and sales driven organizations (aka salesforce) have not made a killing with their IaaS products (database.com). And SAP needs to pick up a lot of speed - as with all things in the cloud - Amazon AWS is the benchmark, redshift is the hurdle - and with Infor endorsing it - redshift is noticeably ahead of Hana at this point.
But congratulations to SAP for embracing key cloud capabilities - May 5th 2013 is a historic milestone for the company - but a lot more needs to be addressed to make Hana a full true cloud success story. As mentioned above, key cloud benefits are still not addressed or need to be addressed soon, first and foremost on the licensing side. A promising start nonetheless.
* For the lucky reader that did not have to learn Latin in school...
- terra incognita - unknown land
- nubes incongnitae - unknown clouds