On Mon, 21 Jun 2010 09:44:10 -0700, Mark Yuhas <mark.yu...@paccar.com> wrote:
>Replacing tape with disk? >Sounds like a great idea - if you have a large enough DASD farm. >DR would be simpler, too - if you have a large communication pipe to >your DR site. > >However, once the powers-that-be see the cost, tape doesn't seem so bad. We thought so too, but then did a complete analysis of where our tape library was. How old are the tapes on the shelves ??? How much does it cost to maintain a tape library ???; floor space, people, hardware... What it came down to for us was: 1) Almost every tape we had was either at or past its expected shelf life. 2) We had no robotics. 3) We are short on raised floor in the data center. A new robotic library with new tapes was about $1.2 million. The various tapeless quotes were $600K - $800K. Everything we have, about 10,000 tapes could fit in half a rack on disk. Almost 600 square feet of raised datacenter floor dropped to 4 Square feet. It's virtual, so no operator intervention at all. Saved a half person and allowed them to do something else. Not sure what, but it isn't loading/unloading/racking/shipping tapes. The data is replicated. We don't have to ship tapes. Saves about $5000 per DR test and two days of shipping. We saved $100K on offsite storage and transport costs. The data is deduped prior to replication. Not a whole lot of bandwidth required to support replication. All-in-all, I think the total cost of the swap was about equal or maybe slightly more than what we would have spent cycling tapes, buying new drives and managing storage (Offsite, onsite, shipping & handling). But we got a lot of benefits from it. We didn't realize how many other things would benefit from this nor the costs we wouldn't incur as a result. We just completed a DR test. We had a couple disk volumes mis-placed. The volume backups on virtual tape were at the hot site, so we just recovered them there. Didn't have to ship a thing. Didn't have to travel. Handled the whole test remotely. I think we calculated the return on investment at slightly over two years. But when you add in all the things we didn't think of, we've probably already paid for it and it hasn't been a year yet. Personally, I feel we should still be managing long-term data on tape. But it's really hard to argue with the results. We don't have to power a bunch of tape drives nor manage the handling and storage of tapes. When the devices reach end-of-life, no matter where we decide to go next, we still don't have to handle tapes nor worry about them falling off a truck somewhere and have that end up in the news. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html