Larry Ellison may open Oracle OpenWorld singing ‘All my Ex’s live in …Hardware’. Doesn’t have quite have the same ring as George Strait’s classic “All my Ex’s live in Texas.” Unfortunately for Oracle, it may not have that song’s lasting appeal either.
With great ballyhoo, Oracle unveiled its Exalogic Elastic Cloud hardware appliance at the 2010 Oracle OpenWorld event. This was being marketed as a ‘cloud-in-a-box’ and aimed squarely at the private cloud market. Results have been mixed. Even with the main spotlight trained on Exalogic the entire show – and with Oracle Fusion taking a back seat.
This year, Oracle is set to take Exadata one step further and unveil a new permutation on the Exadata database machine; this new Frankenstein device goes by the codename “Exalytics” and embeds analytics into the hardware appliance.
To keep all their “Ex’s” in perspective, Exadata is actually a Sun /Oracle appliance that configures storage server hardware, software and an enterprise database. Exalogic is a bundled hardware/software device that features either x86 or SPARC compute nodes, an open standard grid architecture, Oracle middleware and Sun storage and networking capabilities. Now with Exalytics, Oracle is looking to provide an analytics solution to those companies that want analytics capabilities in their everyday business processes.
As they did last year with Exalogic, look for Oracle to attempt to impress with speeds and feeds-and no doubt a high purchase price. At OpenWorld 2010, a single setup of Exalogic was said to have the capability of handling 1 million HTTP requests per second. Impressive number perhaps but enterprises are still figuring out their private versus public cloud environment. And for many large enterprises, at least those that run SAP, there may be a more obvious and strategic choice for real-time analytics.
SAP now offers the HANA in-memory appliance that SAP says can perform ‘real’ real-time analytics processing. SAP’s HANA is an analytics appliance, database and an applications platform all in one. Data is stored in columns (not records), and runs queries under standard SQL, so it can fit under any relational database. Specifically, large customers can get exceptionally fast reporting from usage of HANA. This is because HANA is an in-memory appliance, and all of the data sits within the memory of the application itself. So the calculation engine for analytics also sits in-memory and calculates and crunches numbers in-memory.
For large data workloads many users will find this a plus. SAP already had success with clients using HANA. Chrysler is using HANA to calculate real-time profitability per car sale at dealerships in the U.S. and SAP is targeting use cases in other verticals such as energy, retail and CPG.
From a strategic perspective, what SAP may also achieve is building HANA to be an in-memory data architecture for all its applications, effectively negating the traditional database one day. In other words SAP will try to cut Oracle out of its technology stack with HANA, and based on the technology and technology roadmap, it is quite possible SAP could achieve this.
On the other hand, Oracle is building out a legacy technology, which might be good for the immediate competitive landscape against Teradata and the likes, but SAP seems to have the edge in strategic thinking and is looking toward the long-term situation. Right now Oracle is ignoring this trend but SAP, with HANA, seems to be addressing the future.