Features comparison between Linux SCSI targets
As on April 2009
| SCST | STGT | IET | LIO v3.0 | |
|---|---|---|---|---|
| General | ||||
| Upstream kernel | - | Kernel part | - | - |
| Generic Target Engine | + | + | iSCSI only | Under development |
| Architecture | Kernel only | User space only 1 | Split 2 | Kernel only |
| Stability | + | + | + | Heavy development |
| Performance | ***** | *** | **** | **** 3 |
| Zero-Copy | **** | *** 4 | **** | **** |
| Support for transports without expecting transfer values (parallel SCSI, SAS) | + | - | - | - |
| User Interface | ProcFS | Custom | IOCTL/ProcFS | ConfigFS/IOCTL |
| Major features | ||||
| Target drivers in kernel space | + | + | - | Under development |
| Target drivers in user space | + | + | - | - |
| Backstorage handlers in kernel space | + | - | - | + |
| Backstorage handlers in user space | + | + | - | - |
| Local access to emulated backstorage devices | + | - | - | - |
| Advanced devices visibility management | + | - | - | - |
| Support for Asynchronous Event Notifications (AEN) | + | - | - | - |
| AEN for devices added/removed | + | - | - | - |
| AEN for devices resized | + | - | - | - |
| Bidirectional Commands | + 5 | + | - | - |
| Extended CDB (size >16 bytes) | + 5 | + | - | - |
| Descriptor sense support | + | + | - | - | RESERVE/RELEASE (Windows 2003 clustering) | + | + | + | + | Safe RESERVE/RELEASE implementation according to SCSI requirements 6 | Safe | Safe | Not safe | Not safe | Safe implementation of Task Management commands 7 | Safe | Not safe | Not safe | Not safe |
| Persistent Reservations (Windows 2008 clustering) | - | - | - | Under development |
| ALUA | - | - | - | Under development |
| Failover Clustering | + | ? | + | + |
| SCSI MIBs | - | - | - | + |
| Cluster Storage Integration | - | - | - | VHACS |
| Supported SCSI transports | ||||
| iSCSI | + | + | + | + |
| Fibre Channel | + | - | - | - |
| SRP | + | - | - | - |
| iSER | - | + | - | - |
| Parallel (Wide) SCSI | + | - | - | - |
| SAS | Under development | - | - | - |
| FCoE | Under development | Under development | - | - |
| IBM pSeries Virtual SCSI | - | + | - | - |
| Supported backstorage | ||||
| Kernel side FILEIO | + | - | + | + |
| Kernel side BLOCKIO | + | - | + | + |
| User space side FILEIO | + | + | - | - |
| SCSI Pass-through | + | - | - | Single initiator only, not enforced 8 |
| 4k sectors support in pass-through mode | + | - | - | ? |
| 4k, 2k, 1k and 512 byte sectors emulation in modes, other than pass-through | + | - | - | + |
| CDROM emulation from ISO files | + | + | - | - |
| iSCSI Target | ||||
| Architecture | Split 2 | User space only | Split 2 | Kernel only |
| MC/S | - | - | - | + |
| Max ErrorRecoveryLevel | 0 | 0 | 0 | 2 |
| Bidirectional Commands | + 5 | + | - | - |
| Extended CDB (size >16 bytes) | + 5 | + | - | - |
| Support for Asynchronous Event Notifications (AEN) | + | - | - | - | Safe implementation of connections and sessions reinstatement 9 | Safe | Not safe | Not safe | Not safe |
| iSCSI MIBs | - | - | - | - 10 |
REMARKS:
1. STGT has SCSI target engine in user space and small hooks in the kernel to interact with in-kernel target drivers.
2. All iSCSI management implemented in user space and actual data transfers in kernel space without user space involved.
3. The conclusion made by source code study only. LIO should have performance on the IET level or slightly less, because of more processing overhead.
4. Some zero-copy functionality isn't available from user space. For instance, zero-copy send to a socket.
5. Not well tested, because at the moment there is no backend using this functionality.
6. SCSI requires that if an initiator clears reservation held by another initiator, the reservation holder must be notified about the reservation clearance. Otherwise, several initiators can at the same time change supposed to be protected by the reservation data, which can corrupt them.
7. After a task management command completed and corresponding response was sent to initiator, who sent that task management command, all the affected SCSI commands must get into a state, where they can't affect following after the tasks management response commands from this initiator. This is the safe implementation. The unsafe implementation only marks all the affected SCSI commands as aborted and then immediately send task management response to the initiator. This implementation only guarantees that the initiator will never receive responses from those commands, but it doesn't guarantee that none of those commands will get executed by backstorage *AFTER* any SCSI command, which initiator will send after it received the task management response thinking that all the aborted commands actually fully aborted. This could lead to a data corruption.
8. LIO doesn't emulate all the necessary SCSI host functionality to allow to share a SCSI device in pass-through mode to several initiators. It can only pass SCSI commands from initiator to the SCSI device. This is safe only if there is only one initiator. This limitation isn't enforced anyhow by LIO. You can find more technical information about that in http://www.mail-archive.com/linux-scsi@vger.kernel.org/msg06911.html
9. Connections and sessions reinstatement is, basically, a kind of Task Management command. For instance, open-iscsi uses it as such in eh_target_reset_handler() callback. So, similarly to (7) above, a safe implementation must not accept SCSI commands from new connection/session until all the SCSI commands in being reinstated connection/session get into a state, where they can't affect the new commands.
10. SCSI-MIB (RFC 4455) - not supported; SBE-SCSI-MIB - supported by LIO v2.6, not supported by LIO v3.0; SBE-ISCSI-MIB - supported by LIO v2.6, not supported by LIO v3.0; SBE-IPS-AUTH-MIB - supported by LIO v2.6, not supported by LIO v3.0
P.S. LIO has similar comparison page. It is very much wrong about SCST. I asked that page author, Nicholas Bellinger, to correct the wrong items (see my e-mail), but Nicholas Bellinger refused to do it. Moreover, he blocked my access to the LIO mailing list, preventing me from tell that to the interested people myself. After all our previous discussions and with his skills and experience I can't believe that Nicholas Bellinger doesn't know that SCST is a lot more generic than LIO and has zero-copy in all the same places, where LIO has. So, that comparison page looks like rather a deliberate cheating attempt. Definitely, SCST in all major technical areas is so much superior over LIO, so Nicholas Bellinger started attacking people's perception about SCST trying to inspire them the opposite.