MaxiScale FLEX Software Platform Technology
Peer Set™ Architecture
Benefits
- Always available: Peer Sets™ distribute and replicate file data and metadata, making the MaxiScale FLEX Software Platform resilient to multiple failures and facilitating automatic self-healing
- Runs on standard, low-cost hardware: the Peer Set™ architecture exploits the economics of off-the-shelf server hardware with SATA drives and Ethernet networking
- Scales without bottlenecks: file serving requests are distributed across Peer Sets™ without bottleneck-inducing centralized control points
MaxiScale’s Peer Set™ architecture is a patent-pending technology that delivers linearly scaling performance with each additional cluster node. Peer Set™ instances are a foundation of the MaxiScale FLEX Software Platform and enable a range of functions unavailable in conventional clustered architectures.
A Peer Set™ instance contains multiple members, the number of which is user-selectable based on desired bandwidth and replication characteristics. Each member consists of a low-cost SATA hard drive on a separate physical server node. All file data and metadata within a Peer Set™ is replicated to, and load balanced across, all members.
Learn more in the Distributed Metadata whitepaper.
Single namespace at petabyte scale
Benefits
- Increase file serving capacity without application changes: applications access files with consistent pathnames, avoiding risky namespace updates as capacity increases
- Reduce storage management cost: administrators perform operations once rather than many times including one mount point per server, backups from a single location, and simplified capacity analysis
- Serve more files without namespace fragmentation: increase the number of concurrent files served without file server bottlenecks that force splitting the namespace
The MaxiScale FLEX Software Platform aggregates up to 65,000 Peer Set™ instances into a single namespace, representing hundreds of petabytes at current hard drive densities. Client applications access a file using standard POSIX commands, triggering a direct request to the Peer Sets™ containing the file's metadata and data.
Single disk I/O file serving
Benefits
- Reduce file serving cost by an order of magnitude: single disk I/O reads deliver better performance with far fewer disks compared with typical file servers that require 10 or more operations to retrieve each file
- Predictable web experience across billions of files: give millions of users instant access to their broad range of content requests which are usually difficult to cache
- Deploy file serving that grows seamlessly with web application demands: avoid file serving bottlenecks of conventional sysystems unable to meet a rapidly expanding web-scale business
The MaxiScale Small File Repository (SFR) optimizes web content file serving with low latency access directly from disk. Typical web content files are well under 1 MB in size. Examples include CSS stylesheets, JavaScript libraries, images, and video thumbnails. SFR stores these files contiguously on disk and performs each file operation with a single disk I/O.
