Last Updated on April 27, 2014 by Ruchir Shastri
For an end user, it is always very baffling to choose a right “Cloud” server with so many small hosts popping up who love to market their services as Cloud Enabled. I’ll walk through some important points explaining how we implement VM’s in cloud compared to traditional standalone virtual servers.
We have a completely separate physical space that we use to isolate our cloud infrastructure. The hardware involved with cloud setup includes:
1) Servers – Control server, HyperVisors, Backup Servers – all of these are enterprise class SuperMicro Branded servers.
2) SAN Arrays – We use multiple with 10K SAS drives in RAID 10 array. These have replication enabled and each array in itself is 100% redundant with multiple PSU’s, RAID controllers, RAID 10 disks, etc. All of these units are Dell Equallogic PS series.
3) Switches – Used for public connectivity, storage SAN connectivity, backup network. All of these are different brands of Juniper Switches.
From software perspective, we use OnApp cloud enabling software which manages the logic and intelligence for some amazing cloud features. Our cloud setup has enterprise-class resilience, with multi-layered security, a self-healing architecture, automatic hypervisor failover and disaster recovery supported automated backups.
How it works – Automatic Failover & Hot Migration
We have setup our cloud environment in a way that it is 100% redundant end-to-end. Hence, we can lose half of our entire setup equipment and your service will still continue to operate. There are multiple HyperVisors which host the VM’s and all of these are connected to redundant SAN arrays through fibre channel and 10Ge connectivity. In case one of the HyperVisor (HV) goes down for any reason, VM’s will automatically failover to alternate HV and it continues to work without any connectivity loss. 99% of the Operating systems (except legacy Windows 2003, etc) support hot migration – this means that VM will continue to operate with without reboot in same state even after migrating over to alternate HV (all open programs / services will remain open and operational). In case operating system does not support hot migration, it will be shut down, migrate to another HV and thereafter restarted back. This ensures that VM remains up and working at all times – ensuring 100% uptime guarantee even if HV goes down.
Cloud hosting environment also supports automatic load balancing of selected VM’s. It is just a matter of few clicks to specify VM’s that will participate in load balancing cluster. With such a setup, all incoming traffic is intelligently distributed evenly amongst VM’s in cluster.
Cloud server hosting also supports, easy-to-configure autoscaling for virtual machines. VMs can be configured up and down (adding/removing RAM, CPU and disk resources), and scale VMs out (by cloning the VM) according to configurable thresholds.
For example, you can set up a rule that will add 1,000MB of memory to a VM if RAM usage has been above 90% for the last 10 minutes – but add no more than 5,000MB in 24 hours. Billing is automatically adjusted for autoscaled resources.
Virtual machine backups are also automated with cloud hosting in addition to availability of manual backup and restore functions. The Cloud enables automated/scheduled backups and ad-hoc backups of VM data. These backups in terms of VM snapshots are stored outside of SANs on separate cluster of backup servers, allowing true disaster recovery in case of SAN failure.
We provide pre-built OS templates to build your VM with. There are 25+ supported operating system templates which include various flavours of CentOS, Debian, Ubuntu, OpenSuse, ClearOS along with Windows Server 2003, Windows Server 2008 R2 and the latest Windows Server 2012.
We’re having hundreds of customers hosted on our enterprise class Cloud VPS plans and it has been nothing but a stellar hosting experience with 100% positive customer feedback!