SAP HANA Enterprise Cloud Heralds Arrival of Parallelization Era

With the formal launch this week of the SAP HANA Enterprise Cloud a new era of in-memory computing is about to become a lot more accessible. According to Professor Hasso Plattner, chairman of the SAP supervisory board, HANA in the cloud will give IT organizations access to a Platform that essentially eliminates the need for batch processing in favor of a model where all applications are run in real time.

To accomplish that goal Plattner says takes HANA takes advantage of the parallelization capabilities of Intel multicore processors to allow application performance to for all intents and purposes infinitely scale. The SAP HANA Enterprise Cloud currently consists of 500 servers from IBM, Hewlett-Packard and Cisco that are integrated using high performance switches from Arista Networks. The servers and switches are tightly coupled in what SAP describes as cells, with each cell consisting of 40 servers to provide 100TB of in-memory computing per cell.

In addition to providing a platform-as-a-service (PaaS) environment in the cloud, SAP has developed new management software that automates the provisioning of bare metal servers in a minutes via an Application Programming Interface, while at the same time developing new software that seamlessly integrates the management of servers with the switches from Arista Networks.

Plattner says HANA essentially eliminates the barriers to cloud computing adoption that arise from disk latency issues. Applications running on the SAP HANA cloud now run at performance levels that make the physical location of the HANA servers irrelevant, says Plattner.

While existing applications can benefit from HANA, Plattner says that developers that take advantage of parallelization technologies are going to see the most substantial application performance gains because there will never be a need to make a call out to disk. In addition, applications running on the SAP HANA Enterprise Cloud will be managed by SAP, which will continuously optimize the performance of those applications as part of the service.

In fact, not only will the applications be faster, Plattner says the size of the overall data center environment will be much less because of the compression technologies that allow the HANA platform to process terabytes of data in memory.

The implications of HANA span everything from providing a platform for running interactive mobile computing applications that process queries instantly to eliminating the need to test applications using non-production data.

Next up SAP says it is also working on new approaches to Flash memory storage and server architectures that would all be optimized for in-memory computing.

At the moment, SAP says it has 60 customers running applications on the SAP HANA Enterprise Cloud and about 40 startup vendors building applications for the platform, some of which actually compete directly against existing SAP offerings.

Ultimately, Plattner says that SAP HANA is all about regaining the speed of business innovation, which he says can only really be achieved when all data is running in-memory and the need for batch processing has been completely eliminated once and for all.

Be sure to read the next News Services article: CIR Launches API Allowing Public Access to VA Backlog Related Data