Product detail

MaxiScale FLEX Software Platform

Benefits

  • 1/10th the cost of traditional systems: use inexpensive, non-proprietary hardware with fewer, high-capacity spindles
  • 10X performance increase: serve files under 1 MB in a single disk I/O from every spindle, distributed metadata removes lookup bottlenecks common in centralized systems
  • 100X capacity of current solutions: seamlessly expand file serving capacity to hundreds of petabytes within a single namespace, eliminating the need to point applications at new storage islands

Massively scalable file serving for web applications

The MaxiScale FLEX Software Platform serves files directly from disk up to ten times faster than traditional systems and scales to hundreds of petabytes in a single namespace. The all-software approach scales performance linearly as more nodes are added to the cluster. It runs entirely with standard servers, SATA disks, and Ethernet, dramatically lowering the cost of file serving at Internet scale.

 

View online datasheet. Read about the technology behind the MaxiScale FLEX Software Platform and explore the applications it enables.

Unified Management

Benefits

  • Meet file serving SLAs: analyze end-to-end file serving performance trends to verify the response times to client end-points
  • Automatic, transparent failure recovery: automatically activate spare resources and restore the system to healthy status
  • Reduce storage management cost: add capacity with a single click and ensure proper configuration and load balancing

Manage petabytes of file serving capacity from single view

Unified management cuts the time and cost to operate petabytes of file storage. It improves business outcomes by facilitating service level monitoring, automating capacity growth, and self-recovering from failures.

Learn more by reading the management whitepaper.

Web Workload Testing

Testing Best Practices

  • Utilize open-source tools: achieve maximum transparency and consistent test results with third-party tools. Each test phase should document file staging, simulated user loads, and clear results.
  • Test with large data sets: small data sets are easy to accelerate with caches at the file server, operating system, and even within some applications. Cache acceleration benefits do not apply to a production environment when the live data set is of a much larger scale.
  • Execute multiple iterations: perform tests with varying storage cluster sizes to validate linear scalability of bandwidth, I/O, and capacity. Storage demands grow over time, mandating a file serving solution that increases along all metrics. Systems that perform well at small scale may not deliver expected results at hundreds of terabyte to petabyte capacities.

Linear scaling of file serving bandwidth, I/O, and capacity is critical for Web applications to meet the growing demands of traffic increases from new user content like photos and emails. Architects must evaluate file serving solutions in ways that closely resemble live workloads. Simple file copy tests or legacy benchmarks typically do not represent real scenarios. Live traffic often deals in millions to billions of files, hundreds to thousands of simultaneous requests, at hundreds of terabytes to petabytes of storage. Many users are taking advantage of open-source tools to evaluate Web file serving solutions at production scale without disrupting live applications or developing custom code. Here is a testing methodology that fills this need:

 

These open-source projects and resources can help you get started:

Try MaxiScale Now!

It's easy! Sign up for a risk-free, no obligation MaxiScale software trial today.
Just send us your name and email address so that we can quickly contact you to arrange your trial.

Try MaxiScale Now!

It's easy! Sign up for a risk-free, no obligation MaxiScale software trial today.
Just send us your name and email address so that we can quickly contact you to arrange your trial.