In Wake of CoRAID Demise, Will Open AoE Rise?
-
The Open ATA over Ethernet project has picked up as AoE founding company CoRAID bites the dust. Will moving the protocol to an open standards project allow it to survive where having a vendor steward failed? Only time will tell. The OpenAoE project is very new, their blog is just an announcement at this point and their Twitter feed is about the same. They are working on getting AoE support into XenServer 6.5 now.
-
If AoE was an open standard and the reference implementation was open, I could see it having a good place. Getting good, solid, open source drivers into Linux, BSD, Windows, ESXi, Solaris, etc. will be key. If it was ubiquitous, like iSCSI, then it would be easy to consider it.
-
To clarify a few points, correct some mistakes and give an other point of view...
AoE specs are fully open since day one... :
http://en.wikipedia.org/wiki/ATA_over_EthernetNot available on Coraid web site anymore of course but you can find them here (or search on google) :
http://ftp.alyseo.com/pub/partners/Coraid/Docs/AoE/AoEDescription.pdfLinux driver is also opensource since day one... :
ftp://ftp.alyseo.com/pub/partners/Coraid/Drivers/Linux/Integrated in LInux kernel for years and now available on github under the OpenAoE community :
https://github.com/OpenAoE/aoeBSD and Solaris (opensource version) are outdated :
ftp://ftp.alyseo.com/pub/partners/Coraid/Drivers/
But OpenAoE community will work on it...Some Windows drivers exists and are opensource :
http://winaoe.org/
http://reboot.pro/topic/8168-winvblock/But guess a vendor will need to take over Coraid on Windows and Vmware driver to ensure certification and support for end users.
Many AoE targets implementation are opensource, make your choice :
vblade[1], a userspace daemon that is part of the aoetools package.
kvblade[2], a Linux kernel module.
ggaoed[3], a userspace daemon that takes advantage of Linux-specific performance features.
qaoed[4], a multithreaded userspace daemon.
aoede[5], a userspace daemon with experimental protocol extensions.Vblade is under the OpenAoE community and on github now :
https://github.com/OpenAoE/vbladeSo you may like or not AoE, but IMHO AoE is not dead and has not died with Coraid...
Let's wait and see, it is just the beginning of the story for OpenAoE and new vendors are coming or reborn with an open approach :
http://www.thebrantleycoilecompany.com/ -
@Yacine-Kheddache said:
But guess a vendor will need to take over Coraid on Windows and Vmware driver to ensure certification and support for end users.
VMware at least. However, I think that the current efforts of getting open support natively into XenServer is a good starting point. A for-profit vendor won't necessarily be required if popularity increases, a non-profit foundation can get momentum and if good support for non-VMware virtualization can take root first. Getting Xen and KVM to have rock solid support will go a long way. Then if HyperV gets the support VMware won't be nearly so important (or hard.)
-
@Yacine-Kheddache said:
Let's wait and see, it is just the beginning of the story for OpenAoE and new vendors are coming or reborn with an open approach :
http://www.thebrantleycoilecompany.com/That's the biggest fear is getting the same people involved that were there before with CoRAID. CoRAID was an evil company, actually evil. The involvement of the people associated with that project is the biggest risk to AoE as those people may end up tainting the project.
-
@scottalanmiller said:
VMware at least.
My guess is in the next 5 years vmware will become less and less relevant.
-
@thecreativeone91 said:
@scottalanmiller said:
VMware at least.
My guess is in the next 5 years vmware will become less and less relevant.
I think that's a safe bet since, I would say, that happened already over the last six months. It's already gone from nearly the only thing that NTG recommends and supports to about the only one that never comes up.
-
@scottalanmiller said:
@Yacine-Kheddache said:
But guess a vendor will need to take over Coraid on Windows and Vmware driver to ensure certification and support for end users.
VMware at least. However, I think that the current efforts of getting open support natively into XenServer is a good starting point. A for-profit vendor won't necessarily be required if popularity increases, a non-profit foundation can get momentum and if good support for non-VMware virtualization can take root first. Getting Xen and KVM to have rock solid support will go a long way. Then if HyperV gets the support VMware won't be nearly so important (or hard.)
XCP, XenServer are supported by community (Alyseo contrib) for a while and it is now on behalf of the OpenAoE community.
KVM and Xen just work asis (AoE in kernel), aoetools are opensource also and we released some nice startup scritps (sysv so far but will post new systemd version asap).
Hyperv is next but like Vmware, not easy IMHO to have opensource implementation and gain enterprise trust... will see.
-
@scottalanmiller said:
@Yacine-Kheddache said:
Let's wait and see, it is just the beginning of the story for OpenAoE and new vendors are coming or reborn with an open approach :
http://www.thebrantleycoilecompany.com/That's the biggest fear is getting the same people involved that were there before with CoRAID. CoRAID was an evil company, actually evil. The involvement of the people associated with that project is the biggest risk to AoE as those people may end up tainting the project.
Don't mix and match VCs and Inventors/techies
Guess and hope this will be clarified soon very soon 
-
@mlnews said:
@thecreativeone91 said:
@scottalanmiller said:
VMware at least.
My guess is in the next 5 years vmware will become less and less relevant.
I think that's a safe bet since, I would say, that happened already over the last six months. It's already gone from nearly the only thing that NTG recommends and supports to about the only one that never comes up.
+1 Guys
-
The OpenAoE Project is a great idea, and the inventor of the protocol is involved to supporting the project. Nice work Yacine.
