-
General questions about SAN from new user
Hello,
I’m new into this whole world of shared storage and trying to learn as much an as fast as possible so that I can design a fast and reliable system without making some costly mistakes. After days of researching I think i’ve started to understand how all these things work together but I still have some unclarities about some of the concepts and technical mumbo jumbo.
Keeping it short, I need a fast storage (about 800MB/s read write speeds) and probably available to 2 systems (to begin with). It is my understanding that I would need a Fibre SAN system to achieve this.
In my head I think I’ve already made up my mind to go with CalDigit’s Super Share solution.
My questions are:
1) I keep reading that SAN reads the files block based while a NAS is file based, or something like that…what is up with this block, file thing ? Does it have something to do that with a SAN your files show up as a normal partition to your operating system while with a NAS they show up as…what exactly ?
2) Using a NAS doesn’t let 2 people access and work on the same file at the same time, can this be possible in the first place as long as a NAS doesn’t present itself as a normal disk to the system, this is why one would use a SAN for this type of thing ? Also in big networks do NAS devices function on their own, I mean users connecting to them through their browser and getting the files they need as oppossed to being controlled by a server ?
3) Do all SAN solutions need a separate computer which acts as the server (or metadata controller ? is it the same thing) that controls the SAN software which manages the whole system, can’t I have the computer that i’m editing, compositin etc. on act also as a server and control the SAN (of course in this case it would always have to be turned on for the SAN to function) ? If I have the software installed already on the server can’t I just connect to the server and see the shared storage there as a normal disk drive and start using it, do I also need to have a copy of the software installed in my client machine ?4) In one of Bob Zelin’s tutorial articles that I found on the Cow websites he shows how to design your own SAN ( through Ethernet ). After i’ve read the article I didn’t quite understand the following.
You use the multiple ethernet port card to connect to the SAN storage box. So wouldn’t the storage box need 4 ethernet ports to be able to make this connection or did I understand things wrong and you actually connect to the SAN with the help of the managed switch in which you plug the 4 cables and you also plug into it the SAN storage box and the other editing computers ?
If we do this I still do not understand how will the 4GB link aggregated bandwidth be available to all of the comps because you only have this bandwidth between the server computer and the switch. There were some users that commented that actually the link aggregation arhitecture doesn’t in fact give you 4Gb/s bandwidth if just gives 4 1Gb/s separate sessions and that you would in fact obtain the same result by ditching the switch entirely and connecting directly from the server to the clients if you do not have more clients than ethernet ports on the card.
Also what does MetaLAN actually add to the whole mix ? The management of data needed to keep the bandwidth up ? Another user said it offers volume sharing vs file sharing , what’s that ?
5) In the Final Share solution and similar solutions, software such as MetaLAN/SAN is not needed ? Does the server take on the responsabilites and management of the bandwidth ? And shouldn’t you use the link aggregation on all the computers that are attached to the storage to be able to benefit from the bandwidth ? The shared storage is accesible like a normal disk in these systems ?6) Can I build such a system with a PC server ?
Sorry for all the long questions and thanks in advance for sharing your knowledge.
Mike