Es we reported last fall, Oracle’s updating its Exadata database consolidation platform was a generational change with major architectural changes that turbocharged transaction processing and analytics. That Exadata X8M platform, which was introduced to local customers last fall, will now be available in the Oracle Public Cloud, but with a big new twist. It is tapping the cloud scale to increase the capacity of Exadata for analysis to hold up to 25 PBytes of data.
Oracle Exadata customers have embraced the new platform. In Q1 results released last month, Oracle reported that X8M accounted for the majority of Exadata Database Machine sales and that total Exadata local sales in the quarter increased by 15%. It beat HPE, Dell and IBM Power Servers, which either recorded insignificant growth or decline. The database cloud service (though until now not by the X8M generation) has been an effective feature for new business, with a significant number of Exadata cloud customers new to Oracle.
In summary, Exadata has always been ranked by Oracle as a high-performance, highly scaled, standalone data platform for Oracle database where software and hardware are optimized together. As such, it used premium components like InfiniBand for the backplane, which, although a standard, was not widely adopted. The market for Exadata has been large companies in need of consolidating several databases with requirements for high transaction and analytical performance.
While still addressing the same target market, the Exadata X8M marked a major architectural departure. The headline was the adoption of Persistent memory (PMEM), a new level of storage that can deliver – if configured correctly and in the right form factor – almost the performance of DRAM memory for transaction processing at a significantly lower price point.
But the X8M also has several other important architectural changes, as we pointed out last fall, driven by embracing industrial formal and de facto standards. To summarize, the first is a change in internal network that directly affects performance. The X8M moves Exadata away from InfiniBand, a backplane that originally offered greater bandwidth, to 100 GbE Ethernet. It not only allows customers to use industry standard networks with Exadata, but also opens up another high-performance protocol – Remote Access to Direct Memory (RDMA) that the industry is now embracing over Ethernet with RDMA over Converged Ethernet (RoCE) protocol. It allows much higher performance thanks to, as the name suggests, RDMAs direct access from computing to memory that bypasses the entire stack of OS, IO and network software. This can be critical, especially for complex IOPS-intensive analytical issues. Next is the shift of hypervisor from Xen to KVM, which is a far more popular implementation in the cloud.
And it’s in analytics that the new generation of Exadata Cloud Service adds an extra edge over local X8M deployments. Traditionally, Exadata customers had a choice of capacity by sizing racks, starting with 1/8 racks and moving on to ¼, ½ and full racks. For most customers, the sweet spot has been ½ – ½ rack sizes, which include two computer nodes and three storage servers in ¼ rack. With the X8M, these upper limits are increased to 32 computers and 64 storage servers, effectively increasing maximum storage space to 25 PBytes through compression. It places the Exadata X8M in the data lake realm, a fact that Oracle leverages with its support for multiple data models, from relational to spatial, graph and JSON.
The new Exadata X8M Cloud Service is available, both in Oracle Public Cloud and in hybrid / private cloud Oracle Exadata Cloud @ Customer and Dedicated Region Cloud @ customer. Given Exadata’s high-performance design point, Oracle points to its own benchmarks, which show it better than cloud databases such as Oracle Database on Amazon RDSand with Exadata Cloud @ Customer claims similar benefits over RDS AWS outposts. And given how the Exadata Cloud Service scales computing and storage independently, Oracle claims that Exadata’s high performance and resilience make better use of resources to lower customers’ overall cloud costs because they can perform tasks faster with smaller configurations. To drive this point home, we are waiting for the day when Oracle expands Autonomous database, which uses machine learning to further optimize computing, for the Exadata X8M generation of public cloud services.