NitroSell delivers a fully managed, resilient web service to the retailer, designed to offer global coverage for eCommerce. The service runs on clusters of servers, distributed geographically to achieve maximum coverage and minimum processing time. Each server is securely housed, operating with independent, redundant and high-capacity power supplies and Internet connections. Any server is capable of providing the web service and only one server is required to do so.

The server clusters are used for managed services, not data or network hungry telecommunication or telebroadcasting services. Each server is connected to multiple network providers to offer multiple redundancies in information flow between them and the eCommerce shoppers. Fiber carriers enter the server clusters at disparate access points to protect against complete network failure. Proactive network management tools continually monitor and reconfigure network topology in real-time, based on route efficiency and end-user performance, ensuring the fastest and most reliable network connections.

The server cluster’s HVAC (Heating Ventilation Air Conditioning) systems are N+1 redundant, ensuring that redundant systems can immediately come online in the event of an HVAC system failure. All air is circulated and filtered every 90 seconds to remove dust and contaminates. There is an advanced fire suppression system designed to stop fires from spreading in the unlikely event one should occur. All cables are securely tied down with cable racks suspended from ceilings providing dual routes for all cables.

In the unlikely event of a total utility power outage, all server cluster power systems are designed to run uninterrupted with all servers being fed conditioned UPS (Uninterruptible Power Supply) power. The UPS power subsystem is N+1 redundant, with instantaneous failover if primary UPS fails. For extended utility power outages, routinely tested, on-site diesel generators can run indefinitely.

The server clusters use only fully redundant, enterprise-class routing equipment. All routing equipment is housed in a secured core routing room fed by its own redundant power supply. Fiber carriers enter facilities as disparate access points to guard against service failure.

To ensure a high cache hit rate for cacheable content, such as images or documentation, web requests are directed to specialist servers most likely to contain the requested content. There are more than 20,000 of these worldwide. Cache coherence features ensure that the content served to end users is always fresh. Content that has been marked as uncacheable, such as dynamically generated pages that need to be frequently synchronized, is retrieved from the origin server clusters for each request, as appropriate. Trips to the origin server clusters to refresh and retrieve content are optimized via use of persistent connections, thereby reducing latency through reuse of established connections. Additional connection optimization techniques tune the parameters that govern server-to-server connections, as well as reduce the number of long-haul round trips required to retrieve content embedded in web application pages.

A dynamic site accelerator service also leverages performance enhancing compression technology. Compressing content before it is retrieved from the origin servers is especially effective at reducing transfer times for HTML content to users on slower connections; because application pages are often light on graphics, this technique can be particularly effective. This and other specialist enhancements provide improved end-to-end performance, independent of the shopper and retailer’s internet service providers.

The dynamic site accelerator service provides capabilities for accelerating the delivery of completely dynamic or transactional content, which typically involves an origin server access on every request. These capabilities are delivered through the implementation of path and connection optimization functionality, which includes three key technologies; performance enhancement, pre-fetching of data and transport protocol optimization.

Network edge storage ensures fast transfer of content to specialist edge servers by replicating content to multiple locations and by using a global traffic management system to select the optimal location from which to retrieve content. Replication brings content closer to the edge servers, while global traffic management directs the edge servers to the best storage site for fast content retrieval. This approach provides a significant improvement over content retrieval from a single origin site.
Network edge storage provides terabytes of storage capacity ensuring scalability with customers’ growing storage needs. In addition, mirroring content to multiple locations around the Internet ensures effective scaling with geographic diversity.

Redundancy and fault-tolerance are built into every component of network edge storage; redundant server and network configurations with automatic failover, connectivity to multiple ISP’s, highly-available storage hardware, content mirrored to multiple locations and global traffic management. As a result, WebStore content is always available.

Since NitroSell manage your entire eCommerce solution, equipment, bandwidth, maintenance, upgrades, security, reliability, availability of the network and continued shopper access are worries of the past.