SAP Optimizes BW for HANA, Readies BW for the AWS Cloud
It’s official: SAP BW is coming to HANA, SAP’s in-memory database platform. BW is also coming to the cloud -- in this case, Amazon’s unstoppable AWS.
- By Steve Swoyer
- September 1, 2016
It’s official: SAP’s venerable Business Information Warehouse (BW, for short) is coming to HANA, its in-memory database platform. Previously, HANA had played host to BW via SAP BW on HANA. Next month, SAP will deliver SAP BW 4/HANA -- a version of BW optimized for its S/4HANA ERP suite.
SAP BW/4HANA: get it?
SAP BW on HANA isn’t a kludge, but it is an uneasy marriage of old and new technology paradigms. BW was designed and optimized for an older compute and storage model, one in which data volumes, memory resources, and storage capacities were still comparatively small.
More specifically, BW was designed to run against conventional RDBMS platforms -- including DB2, Oracle, SQL Server, and Sybase Adaptive Server Enterprise-- that wrote data to and read data from physical disk. In-memory database designs weren’t unknown in this paradigm; they were, however, highly specialized. For most potential adopters, an in-memory database was prohibitively expensive.
HANA is an in-memory database platform. It doesn't just run entirely in physical memory (RAM) but is designed to exploit all of the memory in a system -- including the different types of on-chip caches used by Intel and AMD CPUs. These consist of the Level-1 (L1), Level-2 (L2), and Level-3 (L3) caches that are integrated into the CPU package itself.
On-chip caches range in size from 32 KB (for L1, per core) to 8 MB or more (for L3, shared among all cores). HANA tries to maximize its use of on-chip caching because the CPU can both write data to and read data from on-chip cache much more quickly than to physical RAM. For example, HANA will basically “pin” chunks of data or queries in the CPU’s on-chip cache. Instead of having to fetch and re-fetch data or queries from slower physical memory, the CPU can retrieve it from the much faster L1, L2, or L3 cache.
With BW/4HANA, however, SAP says it’s delivering much more than an in-memory optimized data warehouse platform for S/4HANA. The new BW/4HANA comes with its own set of HANA-native objects, improved support for the creation and enforcement of sophisticated data “temperature” policies, and a unified management experience. The new BW 4/HANA runs only on the HANA DBMS, unlike its predecessor, which was certified to run against several RDBMSs.
It’s a version of BW for HANA in every way: tightly coupled to the HANA DBMS engine.
“We have some new capabilities in this that are new to BW/4HANA. We’ve redone the [BW] user interface so that it matches up with the HANA development environment. There’s [a] seamless [transition] between the two [environments], because many customers will go backwards and forwards between environments [and] we don’t want to make them think they’re going into two different products,” says Neil McGovern, senior director of product marketing for data warehousing for SAP.
“We’ve added in more sophisticated multi-temperature support so you can distribute data across hot or warm automatically. You can set up the parameters [such that] any data more than a year old gets moved to warm [storage]. You can do it by the age of data, by key value, [by region] so that any customer in the U.S., you can keep hot, customers in rest of world, you move into warm.”
There are two migration paths for existing BW customers. Users of SAP BW on HANA can upgrade directly to SAP BW/4HANA. Customers using BW in tandem with a third-party -- or “legacy,” in McGovern’s language -- RDBMS must first upgrade to SAP BW on HANA.
BW Comes to AWS
There’s one other kind of big change, too, says McGovern: “We’ve also added a Web interface to support the [BW/4HANA] cloud offering.” That’s right: SAP BW is coming to the cloud -- more precisely, to HANA running in the Amazon Web Services (AWS) cloud. “We’re certified [for AWS] and BW/4HANA [on AWS] will be generally available next month,” he continued. “We’ll support other cloud platforms later on, but Amazon really stepped up to the plate to help us with this.”
If anything, McGovern is understating the case. Late this spring, Amazon unveiled “X1,” a monster new service tier for its Elastic Compute Cloud (EC2) cloud service. X1 VM instances can be configured with up to 2 TB of RAM and 128 virtual CPUs. By contrast, VMs in Microsoft's Azure GS-Series -- the previous record-holder -- top out at 32 virtual CPUs.
VM size does matter -- especially for workloads such as HANA, and HANA's in-memory design also helps offset some of the performance constraints endemic to the cloud. Best of all, Amazon’s X1 tier gives potential BW/4HANA adopters a means to cheaply audition that product -- at a fraction of the cost.
Not coincidentally, SAP today also announced that BW/4HANA is certified to run on up to seven nodes (for a total image of up to 14 TB of memory) on AWS’ X1 instances.
McGovern cited SAP BW customer Fairfax Media Pty, which upgraded to SAP BW/4HANA on AWS. “This is a project they intended to do in BW, [but] they did it in BW/4HANA on Amazon Web Services. It took them three months instead of six months and it took them less than an hour to provision their BW system,” he claims.
“Their reports ... running on HANA are about 10x faster than their experience in the past. ... [In addition, they’re achieving a] 50 percent reduction in development time using [HANA’s] simple object sets and deploying on Amazon Web Services.”
Stephen Swoyer is a technology writer with 20 years of experience. His writing has focused on business intelligence, data warehousing, and analytics for almost 15 years. Swoyer has an abiding interest in tech, but he’s particularly intrigued by the thorny people and process problems technology vendors never, ever want to talk about. You can contact him at firstname.lastname@example.org.