Database High Availability Simplified

Oracle RAC in Oracle Cloud with SkyCluster

Oracle Real Application Clusters (RAC) is best-in-class technology for database high availability. By enabling Oracle RAC in Oracle Cloud, SkyCluster unlocks the on-demand agility of the cloud for the most critical applications, while preserving your full control of your databases.


Why SkyCluster

  • Full control of the database and OS with root access
  • 2, 3, or more RAC database nodes
  • Clustering across availability domains
  • Extra-high storage speed with local NVMe SSDs on bare-metal

Flexible Architecture

SkyCluster provides the flexibility for meeting requirements of applications of various sizes and up-time SLAs. Small sized databases will benefit from cost-efficiency of small VM sizes. Large and performance heavy databases will benefit from bare metal instances with local SSDs. Placing database nodes in different fault domains or different availability domains allows achieving up-time SLAs exceeding most on-premises deployments.

Key components of SkyCluster for Oracle Cloud:

  • OCI Instances: Bare Metal or VM
  • Block Storage and/or Local SSD
  • FlashGrid Storage Fabric software
  • FlashGrid Cloud Area Network software
  • Oracle Linux 7
  • Oracle Grid Infrastructure 19c
  • Oracle RAC database engine: 19c, 18c, 12.2, 12.1, 11.2

Typical Configurations

SkyCluster can be deployed in a variety of configurations. Select one that best suits your specific set of HA, performance, capacity, and cost requirements.


Two RAC nodes

The configuration with two database nodes provides tolerance against failure of a single node. Multi-AD deployment recommended for protection against failure of an entire AD.


Three RAC nodes

The configuration with three database nodes provides tolerance against simultaneous failure of 2 nodes. Multi-AD deployment recommended for protection against failure of an entire AD.


2+2 RAC nodes

Placing 2+2 RAC nodes across two ADs allows protection against an AD failure while maintaining AD proximity between database and application servers during database node maintenance.


Ultra-performance with local SSD

Applications that require ultra-high storage read performance will benefit from using local SSDs with 19 GB/s throughput per node and 2.2M IOPS per node. FlashGrid Read-Local Technology eliminates storage read overhead and makes the full throughput available to your database. Bare Metal Dense I/O instances provide up to 51 TB of local SSD capacity per server. FlashGrid Storage Fabric provides mechanisms for ensuring persistence of the data placed on the local SSDs by maintaining a mirrored copy on persistent Block Storage and by fully automating data reconstruction in case of a loss of an entire compute instance with local SSDs. Block Storage volumes can be added in the same cluster for extra capacity.


Extra-large database

Databases with 100’s TB capacity can be deployed using a scale-out architecture with multiple RAC nodes and multiple storage nodes. Each node adds compute, storage, and network capacity to the cluster. Network speed of 50 Gb/s per node, 64 CPU cores per node, and local NVMe SSDs beat most on-premises systems while providing flexibility to resize the resources when actually needed without over-provisioning in advance.

Shared Storage

FlashGrid Storage Fabric turns locally attached disks into shared disks accessible from all nodes in the cluster. The sharing is done at the block level with concurrent access from all nodes.

In 2-node or 3-node clusters each database node has a full copy of user data stored on locally attached Block Storage disks or on local SSDs. FlashGrid Read‑Local™ Technology allows serving all read I/O from the locally attached disks and increases both read and write I/O performance. Read requests avoid the extra network hop, thus reducing the latency and the amount of the network traffic. As a result, more network bandwidth is available for the write I/O traffic.

Advanced Networking

FlashGrid Cloud Area Network™ (CLAN) enables running high-speed clustered applications in public clouds or multi-datacenter environments with the efficiency and control of a Local Area Network.

Network capabilities enabled by FlashGrid CLAN in Oracle Cloud:

  • Each type of traffic has its own virtual LAN with a separate virtual NIC, e.g. fg-pub, fg-priv, fg-storage
  • Negligible performance overhead compared to the raw network
  • Minimum guaranteed bandwidth allocation for each traffic type while accommodating traffic bursts
  • Low latency of RAC interconnect in the presence of large volumes of traffic of other types
  • Multicast support
  • Transparent connectivity across Availability Domains

High Availability

  • No single point of failure
  • 2, 3, or more active database nodes
  • 2-way or 3-way synchronous data mirroring by ASM
  • Rolling patches
  • Multi-AD delivers higher availability than on-premises systems


Flexible Performance

SkyCluster allows you to right-size your cluster at the start and resize later as needed. No need to over-provision in advance for future growth.

  • CPU resources scalable from 4 to 64 cores per node
  • Memory scalable up to 768 GB per node
  • Storage performance up to millions of IOPS


DBA Friendly

  • Familiar tools: RMAN, DataGuard, GoldenGate, OEM
  • Full control of database features and patching
  • Standard database patching process
  • Quick and easy storage expansion
  • FlashGrid technical support with deep Oracle expertise


Security and Control

  • Keep full control of the database software and the OS
  • Cluster VMs deployed in your cloud account
  • FlashGrid has no access to the VMs or data
  • Data at rest encrypted on Block Storage volumes
  • Oracle Transparent Data Encryption (TDE) option available


24/7 Support

FlashGrid support team has deep Oracle and GCP expertise and is ready to help you with the entire infrastructure stack including:

  • FlashGrid software
  • Oracle ASM and Clusterware
  • OCI instances, networking, storage
  • Linux kernel/networking/storage components
  • Performance optimization and troubleshooting
“We had to migrate dozens of mission critical Oracle RAC databases from Exadata to AWS with FlashGrid SkyCluster, and were able to do it without sacrificing performance or availability.”

Rama Kolli, Sr. Director – Database Management, S&P Global Ratings

“FlashGrid SkyCluster provides us all elements needed for running our mission-critical Oracle databases in Azure: storage and networking software, deployment automation, and 24x7 support.”

Jay Wilder, Sr. Director of Software Engineering, Nuance Communications

“With FlashGrid and AWS we can now deploy a new application within two weeks instead of six months, without compromising our availability SLA.”

David Urban, VP Operations, Aria Systems

“SkyCluster for Azure proved to be a reliable solution that is easy to implement and support for us as a systems integrator.”


“With our new virtualized RAC running on FlashGrid architecture we do a 125GB full table scan in 6 seconds. That is an 11x improvement over our previous setup.”

Sam Shiel, Oracle DBA, Simplyhealth

“Implementing Oracle RAC with FlashGrid in AWS is an important milestone in Digital Virgo’s cloud journey.”

Mikolaj Klimek, Deputy Director IT Operations and Infrastructure, Digital Virgo

previous arrow
next arrow