Hi Lisa,
the main points in the AIT vs LTO contest to me are:
1. AIT is a proprietary format dveloped for the then niche market of
digital media. It is true it has faster access times than LTO. this is,
though, just about the only advantage it can count over LTO. AIT-2
cartridges have 50GB native capacity compared to 100GB for native LTO;
AIT-2 can go up to 130GB for compressed capacity while LTO can reach 200GB.
AIT-2 has faster access times because the cartridge is smaller, so, on
average, the head has to go through a shorter tape length than LTO to get
to the first byte of data; but from then on contest is over, as LTO can
sustain transfer rates of 15MB/s in native mode and 30MB/s for compressed
data while AIT-2 runs, respectively, at 6 and 15.6MB/s.
what this means is that when you are transferring big sequential files, as
seems to be your case, LTO will "beat the pants off AIT" for overall
throughput; an analogy could be 3570 vs 3590. 3570 will get to data before
3590, on average, and then lose out on transfer speed. if you're talking
about start/stop and small file transfer then access times are important,
otherwise access time is much less of an issue. even in a situation like
this last one, LTO has a performance advantage that is quite impressive.
anyhow, no one beats 3570's capabilities for start/stop access situations.
Generation 2 LTO is, at the moment, under test and will be out in a few
months with 200GB native media and 30MB/s, or around that mark, native
transfer rate.
2. I don't know of any AIT automated library that can be compared to
3584LTO as to capacity and footprint; you have up to 248TB of native
capacity for the 3584, and you can start out with a base frame with up to
12 drives and up to 28TB of native capacity. AIT libraries, if I remember
correctly, cannot go further than a few TB(4 I think) and a few drives. If
money is a major consideration and you have a homogeneous environment, 3583
would still outpace AIT and cost a lot less than 3584.
3. LTO is an open standard, AIT is proprietary. what this means is that no
one company can control LTO's roadmap and force customer's choices. LTO has
a set roadmap for the next 4/5 years, and if you don't like IBM tape you
just go out and buy HP or STK or whatever and keep using your media. with
AIT you do what Sony tells you to.
4. LTO is SAN ready. LTO drives and libraries have Fiber Channel attachment
and can be put straight into a Storage Area Network. maybe with a GB
Ethernet, your case it seems, this is not an issue but in general it's an
important point. TSM can drive these libraries and move data over the SAN
with a benefit for LAN traffic (ok, not always as we all know... :-)) . AIT
and 3570 are out of the picture here.
hope this helps.
Cordiali saluti
Gianluca Mariani
Tivoli TSM GRT EMEA
Via Sciangai 53, Roma
phones : +39(0)659664598
+393351270554 (mobile)
gianluca_mariani AT it.ibm DOT com
Lisa Cabanas
<CABANL AT MODOT DOT NET To: ADSM-L AT
VM.MARIST DOT EDU
> cc:
Sent by: "ADSM: Subject: opinion on AIT vs LTO
and 3570 tape technology?
Dist Stor
Manager"
<[email protected]
.EDU>
05/24/2002 03:15
PM
Please respond to
"ADSM: Dist Stor
Manager"
Hello all,
I am trying to determine what would be the best tape solution for our ARAN
(Automatic Road ANalyzer) van data system. Basically, huge amounts of jpg,
mdb and xls data is generated (possibly to the order of 700G per week for
20 weeks, and 20G for the remainder of the weeks). This data will be
staged to a 1.55 TB disk array. The customer wants to be able to access
the current year's data and the preceding year's data. The "Terabyte"
server and two or three processing workstations will be on a private
fiber/GigE network, connected by a Catalyst 3550-12G switch.
Having only actually experience with 3590 and 3570 drives, I was hoping
that someone with experience would be able to give me an opinion (nobody on
this list, right ;-) as to which tape technology would best suit their
situation -- small footprint, hold 5-10 TB of data to start, scalable, fast
access, fast read and write. Money is an object, but for this exercise
let's just keep it an abstraction. I had originally thought a 3584 with at
least 3 drives would be a good start, but then I started looking at the
AIT, whose access time beats the pants off of LTO, and thinking that would
be a better solution.
I had briefly thought about HSM, but the enormous number of files and
directories makes it sound like a really bad idea to me. Any thoughts?
I'd also appreciate any ideas about what kind of x86 box would be suited
for the TSM server for this system. My only experience sizing servers is
with enterprises and RS/6000 servers (I can spend a million here and there
for "my" stuff, but this customer has already spend 1.2 million on this
"turnkey" solution, and isn't going to be real happy to hear that for them
to get their stuff to work and have backups and DR, they need to spend an
additional $100,000, but OTOH, that's just deserts for having left IS out
of the loop in the purchase in the first place.)
TIA and have a wonderful Friday!
lisa
|
