Segmented File System™ architecture with no central metadata server	Delivers linear scalability of throughput from a single file system, a single directory, or a single file Scales to up to 1,000s of servers, 100s of PBs of capacity in a single namespace and 100s of gigabytes/sec of aggregate bandwidth and Millions of IOPS of aggregate throughput Scale performance and capacity independently and non-disruptively through online provisioning of servers and storage	Applications run faster Increased cluster or grid utilization Increased accuracy of business decisions No duplication of data and increased storage utilization "Pay as you grow" solution System Administrators can customize a solution based on price/performance constraints
	Fully-integrated scalable Logical Volume Manager	Device discovery Unique cluster-wide human-friendly device naming User-defined volume configurations for increased performance	Simplified management of very large configurations Lower administrative overhead and TCO
	Multi-purpose Journaling Physical File System	Efficient with small or large files and I/Os, with reads or writes and with random or sequential I/Os Fault resilient	Use the same file system solution for all applications including High-Performance and commercial transactional applications Lower administrative overhead and TCO
	Segmented Lock Manager™	Scalable multi-node locking with no central or distributed lock manager Byte-range locking	Easy data sharing among application processes Tight data consistency Simplified application development
	High Availability	Monitors hardware and software elements and detects failures Automatic active-active server-to-server failover and component level failover Segment level mirroring	Highest service availability Rolling software upgrades Zero downtime for planned system maintenance Provide highest data availability and protection
	Partial online operation	File system remains usable at all times - even when portions of namespace are inaccessible	Minimum impact from catastrophic or multiple cascaded failures
	Dynamic load balancing	Enables re-provisioning of servers and storage while file system is online, balancing utilization of resources	Increased cluster efficiency Zero downtime
	Comprehensive management	Simple, easy-to-use, intuitive, standards based Web-based GUI, Scriptable CLI Cluster status and performance viewable on one-screen	Manage complex parallel file system configurations as easily as simple traditional systems Lower administrative overhead and TCO
	Software-only design	Runs on standard server platforms with no special hardware requirements Uses any combination of SAN-attached or direct-attached storage to build a file system	Preserve existing investments while increasing hardware utilization Allow a best-of-breed approach Highest flexibility for building performance and capacity solutions
	Standards compliant	Uses only industry-standard protocols such as TCP/IP and SCSI	Limits risk Leverages existing infrastructure, further reducing TCO
	Efficient on-disk block allocation	4K or 1K block allocation	Efficient disk utilization for both large and small file environments
	Enhanced Data Protection	Global File System Snapshots High speed SAN based backups File Replication	Recover from accidental deletes and enable backups Minimize backup and restore windows Multiple self-healing replicas of files