My experiment as a technical writer for Solarwinds Storage Manager leads me to learning more about IT storage and storage environments. To understand Storage Manager, I must first understand storage. So, I'm reading an informative book, Virtualization Changes Everything, and I'm learning a good bit including some basic information about storage and storage environments. Below is a synopsis of the first couple of chapters followed by a list of terms culled from the book and from Wikipedia.org. If you're new to IT storage like I am, you might want to read through the list of terms before launching into the synopsis.
What is DAS, SAN, and NAS?
Before virtualization, most x86 based servers met their data storage requirements using Direct Attached Storage (DAS) and a local RAID controller. This produced a reliable, low cost storage solution that met availability requirements for most workloads. But alas, it was unable to meet the high-performance and resiliency needed in mission critical applications. And so the SAN and NAS arrays arrived on the scene.
SAN Arrays or SANs provide LUNs to hosts through Fibre Channels. Over time additional SCSI based protocols such as iSCSI and FCoE emerged. SANs provide the highest IO performance for applications. Below is an illustration from the book showing a SAN supporting a physical server infrastructure.
Another shared storage platform is the NAS array which allows massive parallel access to file systems and unstructured datasets. The most common NAS protocols are Network File Systems (NFS) and Server Message Block (formerly known as Common Internet File Systems or CIFS).
Unlike DAS, SAN and NAS arrays provide capability such as:
- Nonstop data services
- High availability for applications
- Storage of massive volumes of data
- Fast and efficient hardware-based backups
- Data replication capabilities required for disaster recovery
- Direct mapping of storage resources to servers providing granular resource utilization and consumption
And then there was Visualization and the Hypervisor
Server virtualization required a new way of implementing datacenters with ew ways of assigning and consuming resources. Shared resources needed to be logically represented, and so the Hypervisor was born,
Hypervisors are designed to be deployed in highly available cluster configurations requiring data to be served on shared, highly available (HA) storage platforms such as a SAN or a NAS array. This fault isolation means the HA mechanism can ensure the restoration of services in case a Hypervisor fails.
The Shared Storage Pool
In the HA clusters, storage resources are pooled so storage capacity is available to all hosts in the cluster and accessible in on-demand provisioning by VI Administrators. Most Hypervisor users prefer storage pools for providing storage in virtual infrastructures.
Most virtual machines need storage to appear as though it were direct-attached storage. Direct attachment can be simulated using virtual disks. Virtual disks are files presented by the Hypervisor to the Virtual Machine (VM) as a virtual hard drive.
The shared pool model enables virtual infrastructure administrators to consume storage resources from the pool without having to interact with the storage management administrators. Storage pools and virtual disks provide the ability to dynamically provision, clone, and migrate the data of a VM.
At this point, I had a started to get a clear picture of how all the acronyms used in Storage Manager fit into the whole storage picture. So I'll stop here and list some of the terms used in this article.
List of Terms
Direct Attached Storage (DAS) with Local RAID controller - met availability requirements for most x86-based workloads. Unable to meet demands of the high-performance and resiliency needed in mission critical applications.
Storage Area Network (SAN) - dedicated network providing access to consolidated, block level data storage and used primarily in making storage devices such as disk arrays.
Network Attached Storage (NAS) - file-level computer data storage connected to a computer network. NAS is often manufactured as a specialized computer built for storing and serving files.
Logical Unit Number (LUN) arrays - number used to identify a logical unit in computer storage.
Fibre Channel (FC) - High-speed network technology used primarily for storage networking. Can run on twisted pair copper wire and fiber-optic cables.
Fibre Channel Protocol (FCP) - Transport Protocol used predominantly to transport SCSI commands over Fibre Channel networks.
Fibre Channel over Ethernet (FCoE) - an encapsulation of Fibre Channel frames over Ethernet networks allowing Fibre Channels to use 10 Gigabit Ethernet networks while preserving the Fibre Channel protocol.
Small Computer System Interface (SCSI) - Standards for physically connecting and transferring data between computers and peripheral devices.
Internet Small Computer System Interface (iSCSI) - IP based storage networking standard linking storage facilities. Carries SCSI commands over IP networks and is used for data transfers over intranets and for managing storage over long distances. Does not require special-purpose cabling like FC, but can run over long distances using existing network infrastructure.
High Availability Cluster (HA Cluster or Failover Clusters) - groups of computers that support server applications and can be reliably utilized with a minimum of down time.
VI Administrator - Virtual Infrastructure Administrator