Oracle Corp. is introducing new SPARC processors to support scale-out cloud computing, just as many of its customers are looking to migrate some of their workloads to hybrid clouds. The new Oracle S7 processors, which are the processing engines on Oracle’s engineered systems, enable fast analytics and data security – supporting both of these chief features with on-chip, built-in acceleration technology.
End-to-end Computing for Hybrid Clouds
End-to-end computing is key to the new S7-based systems, because they support the same Solaris/SPARC workloads that run on the Oracle SPARC M7-based systems, introduced in 2015. For purposes of consistency and compatibility, Oracle is leveraging on-chip acceleration and on-chip encryption from its scale-up M7 processors, and is adapting it for distributed, scale-out computing. The S7 and M7 systems both run applications and databases by running the Oracle Solaris operating system on the SPARC RISC processor architecture. For longtime Oracle customers, these systems are intended to speed performance for Solaris/SPARC workloads running across on-premises and off-premises systems.
Oracle OpenWorld Conference
We expect to see this end-to-end hybrid cloud approach – linking enterprise datacenters to clouds – to be highlighted at next week’s Oracle OpenWorld conference in San Francisco (Sept. 18-22, 2016). By making and selling both types of systems – M7 and S7 – Oracle is working to increase the total available market (TAM) for its engineered systems designed to run SPARC/Solaris workloads.
Compatibility with the scale-up M7 systems is a priority for Oracle, including compatibility with multiple, older releases of the Solaris operating system. For customers, this signals the availability of multiple options for SPARC/Solaris deployments, for both on-premises and off-premises uses.
Building Block for Engineered Systems
The new SPARC S7 processors, code-named “Sonoma,” were announced in 2015, but are now shipping in Oracle’s engineered systems. They have the same feature-set as Oracle’s M7 processors, introduced in 2015. That will be especially useful for customers in the telecommunications, financial services, government and health care markets, where fast, secure communications and high availability are all vital to mission-critical workloads. These market segments value secure systems – and high performance levels – for applications with near-real-time results.
Compared with M7 processors, the S7 processors are smaller in size, designed for use in more compact systems, such as scale-out clusters used for Big Data/analytics, and cloud computing infrastructure. In that light, the S7 release can be seen as part of Oracle’s move to increase its presence in hybrid clouds, combining on-prem and off-prem systems – and to grow the adoption of the Oracle Cloud public cloud.
Transformation of Data Centers
The S7’s high-density chip design reflects the transformation of data center infrastructure to distributed computing based on clusters of small server nodes. Today, many IT organizations are shifting mainstream enterprise workloads from scale-up systems to scale-out clusters of small, 2-socket server nodes.
These changes in deployment reflect a large move to consolidate and simplify workloads in enterprise data centers, while, at the same time, sending a new wave of analytics and cloud computing workloads to run on cloud services. In addition, there are end-to-end workloads spanning the on-prem enterprise data center and cloud data centers operated by service providers.
This is the case for Oracle’s installed base of SPARC/Solaris customers. Some of these sites are consolidating their older Oracle workloads onto fewer on-prem “footprints” for simplified management and maintenance. In many cases, they are choosing to shift some of their older workloads to a private or public cloud. We also note that Oracle already has small SPARC-based servers in its product line, but the S7 processors will make them more powerful – and the embedded acceleration and security features will make them more similar to the M7-based processors in the Oracle engineered systems product line.
At the same time as the S7 launch, Oracle introduced the Oracle MiniCluster S7-2, an engineered system based on S7 processors that provides dense computing in a smaller footprint than was previously possible with clustered M7 systems. The new system can be installed on-premises, or off-premises in the Oracle Cloud. It is a complement to the larger Oracle SuperCluster M7 system, which is positioned as a scalable resource for workload consolidation in enterprise data centers.
In a computing environment that sees frequent changes, the MiniCluster S7-2 supports automated management for provisioning. Oracle is addressing these requirements with a software-based Virtual Assistant “administrative assistant” that supports installation, patching and system management, reducing operational costs. We note here that the Oracle Virtual Assistant does not require Solaris-specific management skills. Just as importantly, the product can be used by a new generation of administrators that prefer GUIs to command-line interfaces (CLIs).
Scale Up and Scale Out
The transformation of the data center is accelerating industry-wide, with options to move workloads and data migration to scale-out systems – clusters of one-socket and two-socket server nodes. Oracle customers can decide to use both scale-up and scale-out systems within their data center – or reaching into public clouds. There are benefits to using each architecture, depending on the customer’s workloads and deployment strategies.
Although Oracle had been criticized in 2010, when it acquired Sun Microsystems, for being late to the cloud, it has since built a broad portfolio of hardware, software and services for multiple cloud computing use-cases. It hosts and offers the Oracle Cloud public cloud, with its x86-based and SPARC-based Oracle systems, and supports hybrid clouds by linking enterprise data centers with Oracle’s public cloud.
Further, the S7, M7 and Oracle Cloud announcements can also be seen as the backdrop for Oracle’s continuing competition with IBM, which has POWER RISC systems for scale-up and scale-out uses, and the IBM SoftLayer public cloud for cloud services. As survivors of the Unix/RISC wars of the 1990s, Oracle and IBM still see close competition in the mission-critical RISC scale-out computing market.
Oracle is expanding its cloud-computing portfolio with the introduction of S7-based engineered systems, and the MiniCluster S7-2. The S7-based systems are binary-compatible with M7 systems, allowing them to run the same database and application workloads. This supports a broader scale-up and scale-out strategy for SPARC/Solaris systems, providing more options for Oracle Cloud and for hybrid cloud deployments.