>Secure and reliable hosting

Heavy Computing deploys Sun Microsystems Sunfire and SuperMicro enterprise servers, due to their very high-end quality, reliability and flexibility of management to assist in critical situations. With remote ILOM (Integrated Lights-Out Management) and BMC (baseboard management controllers) offering IPMI access, all management is possible from remote - even complete reinstallation or rescue via CD/DVD image.

We deploy all our servers in a mirrored configuration - all servers are twinned to a backup system to takeover in case of failure, each running on dual power feeds for redundancy and elimination of single point of failure.

Heavy Computing does not offer any hosting options that allow our clients sites to risk data corruption or downtime: all sites are run on Raid 1 mirrored-disk (HDD and SSD) systems, optimized for high IO performance on the Linux Ext4 filesystem with ECC ram - common memory errors or failed disks will not affect site performance or availability.

Daily snapshot backups to a mirrored servers are the minimal redundancy requirement for all hosting contracts at Heavy Computing - risking failure is not an option. If a node is downed, sites are brought back up in minimal time using most recent snapshots. Increased frequency snapshots are also available, right up to live cross-sync in our high availability (HA) offerings. HA hosting uses the Proxmox Virtual Environment clustering system and DRBD clustered disk replication, allowing instant switchover between nodes in case of outtages.

All customer images have cooperative/shared access to all machine resources if no other images need them - up to 16 2Ghz+ 64 bit cpu cores and 128 GB of ram, and 50000 IO transactions/s or 500MB/s to Raid 1 disk.