According to SAP is the number of SAP HANA deployments still slightly behind expectations. Mid 2015, roughly 7,200 customers have deployed SAP HANA what’s compared to the total numbers of SAP customers (335,000 SAP customers – 2% SAP HANA and 50,000 SAP ERP customers – 14%) is rather small. The question of the day is – WHY do customers not accelerate the adoption of SAP HANA, especially since every customer knows that by 2025 all traditional RDBMS systems (e.g. Oracle, DB2, SQL Server) will not be supported anymore and certain SAP S4/HANA functionality is only available on SAP HANA.
- Talking to customers in the last years and asking them about the delay in SAP HANA adoption reveals, that many are still struggling to find a compelling business justification for the steep investments necessary to deploy SAP HANA. Just the fact that SAP HANA will provide a 10x to 20x better performance is not a business justification.
- The initial investments in software, hardware and licenses are very steep and with the lack of business justification makes it even harder for CIOs and IT application/SAP manager tough to find the budget to get the SAP HANA projects going.
Now Pure Storage provides a solution that can help customers start their SAP HANA deployment with up to 75% less cost in hardware, software and license, but provides almost HANA-like performance. This will help customers
- Get to SAP HANA deployments faster
- Start with an SAP HANA environment that has almost the same performance characteristics as pure HANA in-memory
- Helps developing SAP HANA use cases without the upfront investments
- Gets compelling results fast
That sounds almost too good to be true. Dramatically reduced cost with almost the same performance? But it is true.
We use a feature called SAP HANA Dynamic Tiering (DT), what is an SAP HANA feature, but needs an additional license (~€10,000 per 1TB). The license for DT can almost be neglected compared to the SAP HANA license that can go up to €159,000 for the SAP HANA enterprise license.
Now in a nutshell, SAP HANA DT provides you the capability to offload certain SAP HANA tables that are not necessarily needed in-memory, due to its ‘warm’ data temperature, onto the storage system. It is quite easy to select the table in SAP HANA studio and SAP HANA will automatically offload the table, plus data onto the storage system. If you run a query against the SAP HANA system, it will fetch data from the underlying DT and provide the results back. It is therefore seamless and simple. I admit the word ‘dynamic’ is a bit misleading, since in my understanding ‘dynamic’ would imply that SAP HANA will determine automatically/dynamically what data can be offloaded and which needs to stay as ‘hot’ data in-memory, but talking to SAP’s HANA product team this ‘dynamic’ capability will be available in one of the next SAP HANA Service Packs. Hope so one rather than later.
Now we at Pure Storage actually tested the performance of the entire SAP HANA system when off loading 25%, 50% and even 75% and the results are astonishing. Of course we had to compare the results to an SAP conducted benchmarkiv that compares Oracle performance with SAP HANA (100% in memory).This test was done by SAP to demonstrate the drastic performance improvements of SAP HANA compared to traditional RDBMS databases. On average the benchmarks shows a 13.9x improvement for SAP HANA. This is also our solution baseline.
The detailed results are shown in the table, but it shows that SAP HANA with 25% of data in memory and 75% of data on Pure Storage the performance is still 10.42x faster than an Oracle system. If you would translate this into query performance then ….
- Oracle would run 100sec.
- SAP HANA DT 100% in memory 7.19sec.
- SAP HANA DT 75% in memory/25% on Pure Storage 8.38sec.
- SAP HANA DT 50% in memory/50% on Pure Storage 8.88sec.
- SAP HANA DT 25% in memory/75% on Pure Storage 9.59sec.
Of course these results are only possible with Pure’s all-flash-arrays.
Performance is only one dimension, the other dimension is the cost savings. The table shows nicely the associated cost savings in software, hardware and licenses when tiering the data from memory onto Pure Storage.
- SAP HANA DT 100% in memory no savings
- SAP HANA DT 75% in memory/25% on Pure Storage ~$757,000 savings
- SAP HANA DT 50% in memory/50% on Pure Storage ~$1,480,000 savings
- SAP HANA DT 25% in memory/75% on Pure Storage ~$2,270,000 savings
Putting both results together, then customers can benefit from dramatically lower SAP HANA cost, but can expect almost HANA-like performance. If you would like to see a demo about SAP HANA DT and the great performance results, let us know.
If you would like to read the results in more details, follow the link to our SAP HANA DT White Paper
SAP HANA DT – White Paper
If you are also interested in the deployment guide and want more details about the setup, configuration of the SAP HANA system and read about best-practices, download the following white paper.
Should you have further questions, please reach out to me or to your local Pure Storage representative.