Get Connected FriendFeed  Twitter Link Blog RSS Feeds |
For most customers, a single Titan has the performance, scalability and availability required to meet their NAS requirements, but for environments with the highest levels of uptime and availability, multiple Titans can clustered together to achieve enterprise level high availability.
Titan's N-Way High Availability Clustering option is a software licensed feature that is implemented in an active/active configuration allowing multiple Titans in a cluster to be actively used for file serving, getting the most efficient use of the systems while proving the highest levels of availability. This is much more efficient than active/passive or active/standby configurations, where a server is idle until a failure occurs, providing only half of the purchased performance.
Titans configured in an N-Way high availability cluster provide true transactional level protection for data even it has not yet been written to disk. Data being written to a Titan is immediately stored in battery-backed NVRAM with up to 72 hours of protection in stand-alone or clustered configurations. In the event of a power failure or loss of access to the storage subsystem, the data is protected within NVRAM and can be played out when power or the storage access is restored. In addition, in an N-Way high availability cluster, the Titans mirror their NVRAM between each other using a high speed inter-cluster network over a 10 Gigabit Ethernet transport. This provides data protection even if a catastrophic event damages one of the Titans, preventing data loss even under these worst case scenarios. This along with Titan's data replication and snapshot capabilities ensure that data is protected and data access is available to users.
Customers can scale BlueArc Titan to 8 nodes. All Titan servers in a cluster share a common pool of storage to best optimize capacity utilization, offer ultimate system availability and achieve new levels of performance as user client and application workloads increase over time. These new cluster enhancements add more control over file system scalability, data migration and placement, project based distribution and workload balancing with support for up to 64 virtual servers. Additionally, implementing clustering allows for the use of two additional software features; Dynamic Read Caching to improve performance scalability and Cluster Name Space (CNS) to provide a virtual global namespace. BlueArc clusters provide customers with the enhancements needed to continue to scale their storage infrastructure in a way that accommodates growth and provide flexibility around ever changing business requirements.
Titan servers are designed for availability. Their hardware-based architecture makes Titan inherently more reliable with less software vulnerabilities than traditional software-based NAS servers. Also, as each function has distributed memory and state engines using many Field Programmable Gate Arrays (FPGAs), processes are isolated and not likely to interfere with each other, unlike software-based NAS servers with central CPUs and memory. As these servers get inundated with requests, processes get starved for resources (CPU, memory, bus bandwidth, etc.) which can cause the system to slow down or even crash.
Cluster Name Space can be used in conjunction with N-Way High Availability (HA) Clustering to provide a unified name space across all nodes in the cluster for simplified data accessibility. Cluster Name Space is optional software license that can be used to virtualize the Titans in a cluster, and provide unified global namespace to all data in the cluster. Customers can use CNS to create a single and persistent virtual file system presentation layer for all data and mount points on the cluster that can scale to 4 PB. This allows for better data management and resource utilization over time while eliminating any disruptive changes to client and application server level access and performance. Unlike some clusters that have separate mounts, each with different data sets, CNS provides access to all the data from any mount point to any Titan in the cluster. All users see is large single virtualized file system, called the CNS tree, with all real file systems residing as sub-directories to this tree structure. Administrators no longer have to provide separate mount points for each node of the cluster for separate data sets, rather users can access all data from any Titan, simplifying user access and reducing file server and data management tasks.
Dynamic Read Caching can be used in conjunction with N-Way High Availability (HA) clustering. This new optional software feature can be added to the BlueArc cluster offerings and introduces a way to help customers optimize workflow related data access and processing capabilities. Cluster Read Caching is a high performance software option designed to scale NFS read-based workload profiles for client users and applications that need shared accessibility and scalable performance into a common shared data set. The read caching approach dynamically and transparently distributes and caches where needed one or more designated data sets across individual Titan Servers within a cluster. The new option has intelligent policies that automatically distribute read-only data sets based on historical client and application access patterns. This new feature allows clustered Titan Servers to automatically distribute customer selected read-only data sets while retaining the level of coherency needed to create a dramatic impact on application speed and productivity.
Careers | Contact BlueArc | Privacy Policy | Terms of Use | Site Map