Increase media access and server scalability while reducing latency and TCO
The growth in Cloud computing is driving new innovations to enable a wider array of supported applications. Email server, interactive databases and transaction processing applications require increased media access, application failover, reduced latency, and greater scalability. Technology is poised to deliver on these requirements while maintaining focus on total cost of ownership (TCO). Two areas of technology that are addressing the latency and cost requirements of the Cloud are virtualization and High-availability clustering.
Small to medium-sized business (SMBs) with local applications can leverage these innovations on a much smaller scale. System failover and reduced latency can mean the difference between a small business’s success or failure. However, to achieve the cleanest access to application data without additional protocol overhead and buffering, the connectivity solution should reside in the direct attach storage interconnect. A low-cost storage-based high-availability solution will help both SMB and Cloud datacenter environments meet their cost and technology requirements.
Storage vendors are working with Microsoft to develop low-cost high-availability solutions for the SMB and Cloud datacenters. These type of solutions will utilize Microsoft’s abilities to manage a clustered server environment through the Windows operating system and a storage vendor’s ability to provide highly scalable interconnect at the storage node.
High availability with SAS RAID Controller Cards
The increased throughput of Serial Attached SCSI (SAS) to 6Gbps and beyond is providing a greater opportunity to fulfill SMB and datacenter high-availability requirements. Direct access to data in a clustered environment via a SAS RAID storage controller offers the optimal solution for the best media access.
With generation over generation of direct attach storage (DAS) interconnect, SAS RAID controllers offer a high level of performance and RAID data protection. As an example, these feature-rich products could use x4 SAS connectors to attach to backplanes and enclosures (see Figure 1). With the shift to the mini-SAS connector going forward, the high-availability environment will benefit from a denser interconnect, pairing SAS RAID controllers on different servers together with external storage.
Figure 1 High-Availability with SAS RAID (LSI MegaRAID®) Controller in the Cloud Datacenter
Architecturally, SAS RAID controllers have the least amount of protocol overhead between the clustered servers and data; keeping consistent a chipset/ PCI bus/ controller Firmware/ SAS protocol IO model. This eliminates the need for additional Ethernet protocol pathways and hardware for virtualized environments. The end result is a huge reduction in application latency and TCO.
2011 Microsoft Build Conference Demonstration- Cluster in a Box
In order to simplify and expedite the implementation of High Availability DAS, a “Cluster in a Box” solution was unveiled at the 2011 Microsoft Build Conference (see Figure 2). The “Cluster in a box” system features two distinct separate servers in a single chassis. Each system can run its own applications while maintaining a single shared storage node. If one of the servers has a failure at the motherboard level or within the OS, the application will fail over to the second server. The second server will continue to run that application until the first server is operational again.
Figure 2 Storage Optimized High-availability Server shown at the Microsoft Build Conference
The “Cluster in a Box” system operates the same way that a 2-server cluster would operate on 2 separate systems. Two SAS RAID controllers furnish the connectivity between server motherboards as well as connectivity to the storage expander via a transition boards. The transition boards connect to the storage expander, which fan out to the 24 drive bays. Each transition board also provides and miniSAS connector that runs external to the enclosure for storage expansion.
