What is an Eva or enterprise virtual array (EVA) And how does this storage technology relate to SAN technology?
If a SAN is part of a data center (should be,) How does the power management rolefit in to data centers in a SAN Environment? Is there any problems with implementation or are the two technologies separate?
How does this storage works NAS work with SAN implemented? Can it serve the role of a SAN? This is not quite clear to me?
Any benefits of using a SAN at home?
Is using a NAS really necessary when you can just share drives from your windows PC over simple file sharing?
I can share files with mac OS-X and do pretty well with that. I am able to pretty much share files to up to 10 computers on my network with Apple file sharing protocol. Under Samba, I think I can let as many people connect to it as the system is willing to handle or max load of the server. Is there any real benefit of having a stand alone Linux or UNIX based NAS?
How does virtualization effect my SAN?
How does cloud computing effect my SAN?
Is a SAN more likely to be a security problem then other types of file systems and protocols?
What does Storage virtualization and the implementing of that effect my SAN and my Network overhead?
I have been looking at how SANS are becoming ever more virtual and need more managed approach to a SAN then just a simple RAID will allow since the size of these disks, allow more error's even 2 errors per disk. there is stuff like raid ten, But then there are other problems that show there ugly head. Like the amount of data corruption that is happening on Memory chips.
Most of these memory chips (dim-ms) are having some really bad error rates. What can be done about that? And what can be done about the Disk Error rate that is happening more and more often. The Google disk test study said that there is an alarming rate of errors. It did not relate witch manufacturer had the most, but only that there is allot more then the statistics really show.
Also, allot of this was for consumer level hard drives. I wonder how much better the Pro series drives really fare. After all it really is all about data. And corruption is corruption. Even if it is on consumer level drives, there should be the same or better standards for BER's to occur.
Considering How SAN's really need more management to really eek out there potential, How does a solution like this fair for SANS and is it really worth it for data centers and where does the benefit end after that?
Should we stick to old disk with lower density's to make sure that we have less problems with data corruption?
You said that you have a SAN set up. What software are you using and How did you get it?
Are there any open source alternatives to SAN's?
Oracle Database Security:
Cost-Effective Data Leak Prevention Starts at the Source
I spoke with Frank at PCC Sylvania in my Network admin class and I am taking again. He described a SAN being a something akin to a CMOS integrated OS with gigabit Ethernet ports connecting to a server using it as a real drive (C:\My computer) instead of a network Share (\my_computer\winxp\etc)
So this would use something like the iSCSI protocol over Ethernet, Infiniband, Fiber channel, etc...
but from a boot point of view that would mean that there would have to be some kind of PIXIE boot server in the BIOS or a mini Linux distro like ASUS uses to serve as a means to get the connection started in the first place to serve as a way to get data from the SAN to the server and then by the time the server os boots, it uses the mini Linux bios as a drive via a a driver in the OS that looks at the information being served akin to what a USB boot drive would be in a sense.
If this is true, then really a NAS can serve the role of a SAN provided the NAS is configurable to run the SAN protocol if the hardware is just a basic computer capable of running a Linux or UNIX operating system like BSD.
A good drawing of this is this one from rackspace.com
the neat thing about this is that you can actually make the storage anything you want. from drive size to space to speed. The OS can leave tasks like fragmentation and other disk utility tasks up to the SAN and leave the CPU doing other important things.
In a sense, it is a NAS. But the key that I see here is independent configurability.
I have not had much experience using much NAS software. However, I have never seen any NAS offer anything of what a SAN can really pull off in its; most basic functions.
Also the SAN uses the ISCSI protocol. It allows you to do some very low level formating of disks and commands that only the SCSI command set provides.
As far as to my understanding the disks in the SAN array can be almost any connection topology. SCSI would be preferred. However SAS & even SATA or IDE and even SSD's could be used to increase flexibility of storage. Storage no longer has any limits and I think that is what a SAN really gives you.
Under a SAN, SSD's make allot more sense and the benefits improve drastically. Since the cost of SSD storage is so high, utilizing it to it's full potential can only be done with a SAN. Unused parts can be used. Parts that are full can be served by short stroked 15k HDD's until the system can make more room and allocate space on the limited space of the SSD.
I would really like to see some real SAN software in action. I know that there are allot of cards out there that offload such tasks from the SAN (like the implementation of the iSCSI protocol and the actual transferring of data over the fabric like TCP/IP offload engines) making the Server and the SAN do less work computing the data and more work doing what they were made for opening up room for more speed and serving more clients on both ends. These special offload engines are make the use of a SAN as effortless as plugging in a flash drive with the speed of a SSD.
In most HDD arrays, there is a limited amount of cache built in to the controller and the drives them selves. in the instance of a SAN, If the SAN has 128 GB of memory, then providing the SAN software is smart enough or capable enough, to use that space for frequent access information, it can sever as a cheap SSD with I/O only being limited by the Connection fabric. There is also the added insurance of ECC and parity to the mix. Often a feater that is missing in commercial offerings such as the RAM SAN and more commercial offerings such as those of Texas instruments.
Overall a SAN's computing power can also be doubled if needs merit it to be. This makes it ideal to deal with database and large data center operations that may run into computing power overhead bottlenecks. Doing file searches like looking for Torrents in a database can be significantly speed up if the SAN takes care of the indexing for the Server OS. This leads to much higher I/O when handling such things such as databases query's and searches. In essence the server does what the server does best. It serves. And leads the DiskOperatingSystem do what it does best. Actually handle the part that it was made for.
Here is a ripe question. How secure is a SAN? If it is based on TCP/IP, then you could latterly run it over the Internet if you wanted to over a VPN tunnel or using SSH (both being accelerated by the network cards Offload engines if so chosen) and have pretty good security.
This type of traffic could also run an encryption algorithm is so desired to handle more sensitive needs.
The only real issue of a SAN is Data corruption. The SAN itself may suffer from small bit-errors that could lead to catastrophic issues within the application that they are used in. This storage problem is more of a issue with the HDD media itself and not with the SAN itself. So in the interest in keeping a real life approach to things. 2 SAN's in operation working independent of each other with respect to there backups. If there is an issue, Reloading the data from a prior back-up may lead to the problem or fix it altogether. A SAN is still not immune to the laws of safe data practices. Keeping back ups is highly recommended in my opinion.
Mickey (Salesh Mishra)
January 8, 2010
