| 3 | | High availability refers to a system and associated service implementation that is continuously operational for a long period of time. Whole-system replication is a conventional way to increase system availability ---- |
| 4 | | once a primary machine fails, the running applications are migrated and resumed on backup machine(s). |
| 5 | | ---- |
| 6 | | However, there are several limitations that make this method unattractive for deployment: it needs specialized hardware and software, which are usually expensive. Additionally, such a system may also require complex customized configurations, which are difficult to manage and maintain. |
| | 3 | High availability refers to a system and associated service implementation that is continuously operational for a long period of time. Whole-system replication is a conventional way to increase system availability: once a primary machine fails, the running applications are migrated and resumed on backup machine(s). However, there are several limitations that make this method unattractive for deployment: it needs specialized hardware and software, which are usually expensive. Additionally, such a system may also require complex customized configurations, which are difficult to manage and maintain. |
| 8 | | These limitations are efficiently overcome by virtualization: all applications now run on a virtual machine (VM). Thus, whole-system replication can easily and efficiently be implemented -- a copy of the whole VM is continuously checkpointed and saved on a backup machine. As VMs are totally hardware-independent, the cost is much lower compared to the hardware expenses in traditional high availability solutions. Besides, virtualization technology can facilitate the management of multiple VMs on a single physical machine. With virtual machine monitors (VMM), the service applications are separated from physical machines, providing increased flexibility and improved performance. |
| | 5 | These limitations are efficiently overcome by virtualization: all applications now run on a virtual machine (VM). Thus, whole-system replication can easily and efficiently be implemented: a copy of the whole VM is continuously checkpointed and saved on a backup machine. As VMs are totally hardware-independent, the cost is much lower compared to the hardware expenses in traditional high availability solutions. Besides, virtualization technology can facilitate the management of multiple VMs on a single physical machine. With virtual machine monitors (VMM), the service applications are separated from physical machines, providing increased flexibility and improved performance. |
| 34 | | * [http://xen.org/ The Xen Hypervisor] -- Xen: Open Source Industry Standard for Virtualization |
| 35 | | * [http://nss.cs.ubc.ca/remus/ The Remus project] -- Remus: Transparent High Availability for Xen |
| 36 | | * [http://www.osrg.net/kemari/ The Kemari project] -- A VM Synchronization Mechanism for KVM |
| 37 | | * [http://friends.cs.purdue.edu/dokuwiki/doku.php?id=vnsnap The VNsnap project] -- Snapshots for Virtual Networked Environments |
| | 31 | * [http://xen.org/ The Xen Hypervisor]: Xen: Open Source Industry Standard for Virtualization |
| | 32 | * [http://nss.cs.ubc.ca/remus/ The Remus project]: Remus: Transparent High Availability for Xen |
| | 33 | * [http://www.osrg.net/kemari/ The Kemari project]: A VM Synchronization Mechanism for KVM |
| | 34 | * [http://friends.cs.purdue.edu/dokuwiki/doku.php?id=vnsnap The VNsnap project]: Snapshots for Virtual Networked Environments |
