Hyperconverged Infrastructure (HCI) has quickly gained adoption in many IT environments due to its relative simplicity and the ability to efficiently scale the number of applications for private cloud and virtual environments. As HCIsystems have matured, IT administrators and application owners have grown comfortable with their capabilities, particularly the ability to quickly scale their deployments while maintaining efficiency through the use of a consistent set of integrated management tools.
According to the latest Evaluator Group research of HCI, many companies express concern over the capabilities of their HCI environments, particularly regarding performance and scalability, along with data protection and disaster recovery features. In order to assess the capabilities of NetApp’s HCI system, Evaluator Group was asked to evaluate it in our lab using a variety of typical enterprise workloads and use cases.
In summary, the NetApp HCI system was able to leverage its disaggregated HCI design. Tested as a solution, the system was shown to scale storage performance and capacity independently from computing capabilities. Additionally, the impact of a storage failure was minimal while running a set of common virtual applications. NetApp HCI Test Highlights:
- Performance: Industry leading HCI storage performance (3 compute + 5 storage NetApp HCI)
- Scalability: Ability to linearly scale capacity and performance of storage independently
- Availability: Tested ability to withstand storage node failure with minimal compute impact
- Enterprise QoS: Tested QoS control per volume and per VM (using iSCSI and VVOLs, respectively)
- Always on Data Reduction: All testing and results with data reduction enabled
Enterprise users have rapidly adopted HCI as a means for rapidly deploying new applications using a common set of technologies and management interfaces. The high degree of integration is what enables IT administrators to deploy, scale and manage new application deployments much more efficiently than using older designs, processes and tools.
As HCI has matured, application owners and IT administrators have both gained confidence in these systems’ capabilities to handle a wide variety of applications that were traditionally run on dedicated, highly customized systems.
Scaling Enterprise HCI
NetApp’s HCI solution is different than many first generation HCI systems by utilizing independent resources for compute and storage. This architecture is often referred to as “disaggregated” in that it breaks apart some of the resources used to deliver an HCI platform. By separating resources, NetApp’s HCI system is able to scale compute and storage performance independently, with almost no impact to other resources. Clearly, scaling performance is one advantage of disaggregation, but there are additional advantages including the ability to deliver high availability without impacting application performance.
The ability to insulate applications from each other has been a concern of application owners in shared environments, including for both virtualization and containers. For these reasons, many business-critical applications have resisted the use of shared resources, be that compute or storage. Isolation is one means of achieving predictable performance, although it comes at the price of reduced efficiency. The combination of dedicated storage processing insulates application demands from potential I/O delays for other applications, and the NetApp HCI QoS capabilities provide additional control for administrators and application owners. These enterprise-class features help enable cloud operators in both private and service provider environments to deliver consistent performance to applications regardless of what other workloads are running.
NetApp HCI Overview
NetApp’s HCI system is designed for enterprise environments that require the ability to scale storage performance and capacity independently from computing, in order to match application requirements efficiently. NetApp’s HCI architecture provides enterprise capabilities along with NetApp Data Fabric components to extend usage, data protection and deployment options.
NetApp HCI Features
- Independent Scale-Out – Scale storage performance and capacity independently from compute
- Data Efficiency – In-line de-duplication, compression and thin provisioning increase storage efficiency by 5 – 10x
- Storage Capacity – From 11.5 TB – 1.8 PB raw capacity (50 TB – 5 PB+ usable with data reduction)
- Clustering – Compute scales to 64 nodes per cluster, storage nodes scales from 4 – 40 nodes
- Quality of Service – Integrated QoS provides ability to control I/O to isolate applications
- Data Resiliency – Dual redundant copies of data distributed to all nodes, automated drive rebuilds
- Data Protection – Native snapshot-based backup and restore functionality to object storage via S3 or SWIFT compatible API
- DR & Replication – Synchronous, asynchronous and snapshot replication locally and between remote clusters
- Availability – Automated failover and failback available between a cluster and up to four other clusters
- Data Security – Encryption with 256-bit AES for environments requiring data at rest protection
- Connectivity – Storage connectivity via iSCSI, Fibre Channel, VVOL’s, and container native storage via NetApp Trident
- Deployments – Public and private cloud deployments enterprise environments
- Integration – VMware vCenter, VAAI, VVOLs, SRM/SRA, Microsoft VSS Provider, PowerShell – Integrated with OpenStack, Containers and NKS via Trident framework
NetApp HCI is designed for private cloud and multi-tenant environments supporting a wide range of applications including general purpose VMs, VDI, Oracle, SQL Server, and NoSQL databases on VMware with support for RedHat OpenShift Container platform. The inherent capabilities of NetApp HCI provide data resiliency and protection, along with disaster recovery and other high availability features. To minimize application impact on shared resources, the system provides quality of service features for fine-grained control of performance.
The NetApp HCI system utilizes a disaggregated architecture to enable independently scaling storage and compute capabilities in an integrated HCI appliance. According to Evaluator Group’s continuing survey of enterprise use of hyperconverged systems, many value HCI systems’ ability to be deployed more rapidly than traditional equipment, along with the ability to reduce complexity. However, some IT users express concerns regarding the ability to scale HCI performance and inflexibility of some HCI systems.
The NetApp HCI provides the benefits of a hyperconverged architecture, with increased scale and flexibility features of converged infrastructure, by enabling independent scaling of both storage and computing capabilities. Additionally, the system can be extended to leverage NetApp Data Fabric services, such as ONTAP Select for HCI, to add file and the full set of ONTAP data services to NetApp HCI. For disaster recovery, NetApp HCI may use SnapMirror replication to a NetApp ONTAP system. With ability to scale capacity and storage performance, the NetApp HCI systems offers a greater degree of flexibility than many competing HCI systems.
Additionally, the disaggregated architecture also allows a greater degree of isolation between storage and application demands, which was demonstrated by both the QoS capabilities and the sustained operations during a storage node outage. Many competing HCI solutions require over-provisioning resources, such as using fully redundant clusters, in order to provide continued performance during a node failure. NetApp HCI’s QoS feature allows the system to operate at higher utilization, while still ensuring less important applications do not impact more critical ones.