Hybrid and multi-cloud approaches to high performance computing (HPC) have great potential to reduce costs, increase flexibility, and improve access to cloud-scale compute power. However, concerns about data protection and management continue to inhibit a broader use of these approaches. NetApp Private Storage (NPS) for Cloud, deployed in conjunction with infrastructure partners Amazon, Microsoft, and Equinix, can overcome the obstacles and make both hybrid and multi-cloud approaches a perfect match for several popular use cases.
Perhaps the best way to illustrate this is to look at how a leading research university deployed a hybrid cloud using OpenStack and Amazon Web Services (AWS).
A unique blend of public and private resources
The university initially built its private cloud infrastructure using Red Hat OpenStack Cloud Infrastructure. At the same time, its researchers were using AWS for various projects. However, the university needed an infrastructure solution that would enable the researchers to access the AWS compute resources without also storing their sensitive data in the Amazon cloud.
To accomplish this, the university created a hybrid cloud infrastructure made up of three components – a private cloud within the university, a public cloud powered by AWS, and a NetApp Private Storage solution co-located at an Equinix facility.
This combination provides the university with a secure and robust architecture that supports several distinct use cases, including:
Cloud-based HPC: The university’s HPC center provides compute-intensive IT resources for academic research. The workloads can be highly variable. For example, for real-time weather forecasting, a storm model is combined with ocean tide data to determine a storm’s impact on the tides. The data volumes generated for such forecasting projects is relatively low on normal days, but spikes occur when weather probes detect abnormal activities. When that happens, the solution must be able to scale to ensure that the high volume of data is captured and analyzed in time for use by a variety of government agencies.
With the hybrid cloud solution, compute resources are automatically provisioned and scaled-up to ensure that the data can be captured and analyzed in real-time. The resources are then de-provisioned when no longer needed, eliminating the need to invest in data center resources for peak workloads that would remain idle on most other occasions.
Development and testing: Being a leading research university, the workload fluctuates frequently as new programs are undertaken and new research directions are explored. To support its research initiatives, new modeling, simulation, analysis, and visualization applications must be developed and tested all the time.
In a typical data center, developers rely on IT to set up the systems needed to deploy and test their applications. This process normally takes from weeks to months and requires significant planning and capital. Frequently, the equipment is only used for a short period of time and once the application is moved into production, the dev/test environment is no longer needed.
A hybrid cloud approach greatly simplifies matters. Developers can use self-service capabilities to provision their systems using AWS compute resources and the secure NetApp storage. The virtual server resources are only used for the duration of a development project. Once the project is completed, there is no unused hardware and no ongoing costs for the server infrastructure.
Disaster recovery: The university also needed an affordable secondary site for disaster recovery. The hybrid cloud architecture allows the university to provide disaster recovery without the need to spend money on dedicated facilities and infrastructure that would remain idle most of the time. Instead, the on-premise data is replicated to NetApp systems that are hosted by Equinix and connected to AWS, which allows administrators to quickly spin-up new machine instances in the event of a disaster. During disaster recovery, the users simply redirect their jobs to compute nodes at AWS for transparent access to the replicated data which is housed at the colocation environment.
For more information about how a hybrid cloud with NetApp Private Storage can help with different use cases, download our whitepaper.