This blog is a follow-on to an earlier blog I wrote on Why Scale Out and What’s New. This blog will focus on support for unified scale-out or multiprotocol unified architecture, one of the many new benefits of Data ONTAP 8.1 operating in Cluster-Mode.
First, let’s define some terms.
Two controllers configured together to provide high availability (HA)
The Data ONTAP operating mode that supports interconnection of nodes into a cluster
A physical system consisting of one or more nodes that are interconnected and managed as a single system
A (one) controller that is a member of a cluster
A unified architecture is the ability to support multiple data access protocols concurrently in the same storage system, over a range of different controller and disk storage types with common management, common software, integrated data protection and support for multiple tiers of storage (primary, secondary, and archive/compliance), as well as quality of service, and the ability to act as a front end for 3rd party storage systems
Since introducing the industry's first unified storage system almost 10 years ago, Data ONTAP 8.1 operating in Cluster-Mode now supports the full range of data access protocols. The supported protocols are: iSCSI, Fibre Channel, FCoE, NFS v3, v4, v4.1, including pNFS and SMB 1 and 2. Data replication, management software and storage efficiency features are seamlessly supported across all protocols in Cluster-Mode.
SAN (iSCSI Fibre Channel and FCoE) Support
Data ONTAP provides the ability to create and make available LUNs to attached hosts. Since the cluster consists of multiple controllers or nodes, there will be multiple logical paths to any individual LUN, and the best practice is to configure at least one path per node in the cluster. Asymmetric Logical Unit Access (ALUA) is used on the hosts to make sure that the optimized path to a LUN is selected and made active for data transfer.
NAS (NFS, CIFS (SMB) Support
Data ONTAP can provide a single namespace. NAS clients can access a very large data container using a single NFS mountpoint or CIFS share. Each client only needs to mount a single NFS file system mountpoint or access a single CIFS share, and only the standard NFS and CIFS client code for each operating system is required.
The namespace within Data ONTAP is composed of potentially thousands of volumes junctioned together by the cluster administrator. To the NAS clients, each volume appears as a folder or subdirectory, nested off the root of the NFS file system mountpoint or CIFS share. Volumes can be added at any time and will be immediately available to the clients, with no remount required for visibility to the new storage.
Data ONTAP also supports the concept of multiple, securely partitioned namespaces to accommodate requirements for multi-tenancy or isolation of particular sets of clients or applications.
With Data ONTAP 8, capacity scales from terabytes to tens of petabytes, all transparent to the running applications. Storage is virtualized across as many as 12 HA pairs, managed as a single logical pool of resources and name space. Virtualizing storage across multiple pairs of controllers provides nearly limitless scalability for even the most data-intensive environments, regardless of protocol, SAN or NAS. Data ONTAP 8 offers the first massively scalable unified storage platform providing cost efficiencies, investment protection, configuration flexibility and ease of use.
My next blog in this series will focus on the key Data ONTAP 8 benefit of nondisruptive operations.