By Gregory Kincade:

As data needs continue to grow, the latest direct attach storage from Dell helps seamlessly expand the capacity of 13th-generation Dell PowerEdge servers.

Today’s enterprises are confronted by an explosive growth of business-critical data, fueled by the prolific use of mobile devices, cloud and social media. No longer consisting of only predictable, structured formats such as database tables, data now includes diverse combinations of email, video, audio, spreadsheets, medical records and other semi-structured and unstructured formats.

The shrinking IT budgets of many small and midsize organizations simply cannot keep pace with the growth and complexity of this data. Yet to remain competitive, these organizations must maintain the reliability, availability and security of their data. They need cost-effective, scalable enterprise storage that can handle both the expanding volumes and the increasing diversity of data. Moreover, the storage solution should include a simple-to-manage, powerful storage platform that supports applications and services requiring an efficient, fully redundant infrastructure to maintain high availability.

Next generation of direct attach storage

Direct attach storage (DAS) is well suited for situations in which throughput and price are the most important considerations and the advanced features of network attached storage (NAS) or storage area networks (SANs) are not required. DAS offers performance benefits over NAS and SAN because the server does not need to cross the network for reads and writes. As a result, DAS can efficiently support demanding applications such as email servers and audio and video streaming.

In addition, DAS is designed to be simpler to deploy and maintain than NAS or SAN. NAS and SAN require network planning and the acquisition and installation of switches, routers, cabling and connections. By avoiding these costs and complexities, DAS remains a viable choice for small businesses or for branch offices. Consequently, for many remote and small office deployments, DAS usually enables greater throughput and can be more cost-effective and easier to configure, compared to NAS and SAN.

To deliver substantial, scalable storage capacity, a RAID array can be appended to a daisy chain of DAS units. Enterprise RAID solutions are designed to offer much of the fault tolerance, availability and data protection provided by NAS and SAN with features such as disk, controller and cooling redundancy. When anchored by a well-designed RAID array, DAS can provide a solution that is easy to own and manage. Nevertheless, pure DAS solutions traditionally were not deemed viable when storage needed to be shared or virtualized and scaled.

Now, scalable and dependable storage solutions that incorporate DAS are ready for the spotlight. In concert with innovative 13th-generation Dell PowerEdge servers, Dell has released its next generation of DAS enclosures: the Dell Storage MD1400 and the Dell Storage MD1420 12 Gbps Serial Attached SCSI (SAS) storage enclosures. The MD14xx models are designed to provide storage expansion for the latest PowerEdge servers using the Dell PowerEdge RAID Controller 9 (PERC9) H830 adapter or the Dell 12 Gbps SAS host bus adapter (HBA).

Enhanced reliability and fault tolerance through RAID

The PowerEdge R730xd, PERC9 H830 adapter and MD14xx combine to form an excellent platform for robust Microsoft® Exchange 2013 mailbox servers and storage for Exchange 2013 mailbox databases and transactional logs. The MD14xx is designed to support the resilience of the Exchange solution by eliminating single points of failure with hot-swappable redundant fans, power supplies and Enclosure Management Modules (EMMs).

The PERC9 H830 adapter, which enables management software to recognize connected storage as a single unit, can detect and use redundant paths to drives contained in MD14xx enclosures. Two SAS cables can be connected between the adapter and an enclosure for I/O path redundancy (see figure).

Fig. 1

Redundant path configuration: PERC9 H830 adapter and Dell Storage MD14xx enclosures

Path redundancy enables the PERC9 H830 adapter to withstand failure of a cable or an EMM by switching over to the remaining operational path. When redundant paths exist, the controller automatically balances the I/O load through both paths to each disk drive. This load-balancing feature, which is automatically turned on when redundant paths are detected, helps increase throughput to each drive.

The PERC9 H830 adapter supports the Exchange 2013 recommended RAID levels — RAID-1, RAID-5 and RAID-10 — for system partitions, database files volumes and database log files volumes. The controller is designed to automatically detect and rebuild failed physical disks when a new drive is placed in the slot where the failed drive resided. In addition, if a configured hot spare is present, the controller automatically and transparently tries to use it to rebuild failed physical disks.

If drive security is a priority, the option to use self-encrypted drives (SEDs) with drive-level encryption helps ensure data is secure, even if the drive is removed. Additionally, SEDs support the instant secure erase feature, which is designed to permanently remove data when repurposing or decommissioning drives.

MD14xx enclosures can be managed easily through Dell OpenManage systems management tools, leveraging the Integrated Dell Remote Access Controller (iDRAC) to help simplify and automate essential management tasks across server, storage and networking platforms in multi-hypervisor and OS environments.

The MD14xx models offer the flexibility to mix and scale a variety of nearline SAS, SAS, SED and solid-state drive (SSD) formats within a single enclosure. Up to eight MD14xx enclosures — four per PERC9 port — can be connected to each PowerEdge server, helping simplify the designer’s task of customizing an Exchange 2013 storage solution to fit the desired cost, capacity and IOPS profile.

Cost-effective virtualization through software-defined storage

Software-defined storage (SDS), in which the storage management stack is embedded in the OS, offers a low-cost alternative to traditional RAID storage for efficiently managing and protecting mission-critical data. Microsoft addresses the SDS challenges of building a scalable, resilient and cost-effective storage platform through Storage Spaces, available with the Microsoft® Windows Server® 2012 R2 OS.

Storage Spaces is virtualized storage based on a disk-pooling paradigm. It enables administrators to convert externally connected enclosures of physical disks into storage pools. From the storage pools, administrators can create thinly provisioned virtual disks, which are also known as storage spaces.

Administrators can configure storage pools using Storage Spaces and dynamically add them as needed. This feature not only facilitates storage scalability, but can be used for fault tolerance because administrators can configure spaces with mirroring for clustered Microsoft® Hyper-V® or Microsoft® SQL Server® workloads and parity redundancy for backup and archival workloads. Mirroring allows data corruption to be seamlessly repaired across mirrored volumes.

To support deployments that require an added level of fault tolerance, Storage Spaces can associate each copy of data with a particular Just a Bunch of Disks (JBOD) enclosure. This capability is known as enclosure awareness. With enclosure awareness, if one enclosure fails or goes offline, the data remains available in one or more alternate enclosures. The MD14xx storage enclosures support the SCSI Enclosure Services (SES) necessary for this feature to work properly.

Administrators also can use Storage Spaces with a two-node server cluster for failover, where one or more pools are clustered across multiple nodes within a single cluster. For example, the MD14xx enclosures can be connected to a pair of PowerEdge R730 servers with clustered file server roles; each server has redundant paths via the Dell 12 Gbps SAS HBAs to all the disks in each MD14xx enclosure (see figure).

Fig. 2 

Example of two-node failover cluster configuration: Two servers, each with two HBAs and four storage enclosures

The two-node cluster depicted in the figure is an excellent example of the recommended starting configuration. It supports three-way mirroring and enclosure awareness, helping to simultaneously protect against an enclosure failure and two disk failures across different enclosures.

By using Cluster Shared Volumes (CSVs), administrators can unify storage access into a single namespace for ease of management. All Hyper-V cluster nodes can access a CSV at the same time for direct I/O. This unified namespace allows highly available physical or virtual workloads to transparently fail over to another server if a server failure occurs.

Another Windows Server 2012 R2 capability enabled by storage spaces is automated data tiering. The most frequently accessed data is automatically prioritized to be placed on high-performance 12 Gbps SAS SSDs, and the less frequently accessed data on high-capacity, lower-performance hard disk drives (HDDs). The data is periodically moved to its appropriate location with minimal impact to the running workload.

Storage Spaces includes the ability to automatically rebuild storage spaces using free space in a storage pool. If a physical disk fails, Storage Spaces regenerates the data that belongs to the failed physical disk without any user intervention.

Finally, Storage Spaces can use existing SSDs in the storage pool to create a write-back cache that is tolerant of power failures and buffers small random writes to SSDs before later writing them to HDDs. Small random writes often dominate common enterprise workloads, and they can impact the performance of other data transfers that are taking place. By using SSDs — which excel at random access — for a write-back cache, Storage Spaces helps reduce the latency of the random writes and the performance impact of other data transfers.

Versatile storage expansion

The right storage approach — traditional hardware-based RAID or SDS through Storage Spaces — depends on an organization’s budget and fault tolerance. The intelligence of the PERC9 H830 adapter helps increase reliability and fault tolerance, and the 12 Gbps SAS HBA is well suited for an SDS implementation. Along with 13th-generation PowerEdge servers and MD1400 or MD1420 DAS enclosures, either configuration makes an exceptional building block for the storage infrastructure projects of small and midsize organizations, as well as remote and branch offices of large enterprises.

Learn More

Dell Storage DAS