>Hello all, Is anyone out there doing remote disk(IBM Mainframe) for DR?
If yes how and are you happy with it.

Need to answer some basic questions 

1. What is your RTO (Recovery Time Objective) or the time to be back in 
operation. Obviously it is fairly short for you are thinking about more than 
restoring dumps at some remote site. 

2. What is your desired RPO (Recovery Point Objective) or how up to date 
must the data be at the DR site? This is the most important one to answer. 

     a. If it must be absolutely current, then you are talking about either 
    
         1) Synchronous Mirroring with a distance limitation of not too far and 
most all of the DASD vendors play in this arena. But then in a true disaster 
what are the chances the region you have both computers is "toast".  

         2) For remote DR sites beyond Synchronous distances, you need to  
look at what was called IBM's XRC (now they are into Global "whatevers"). I 
undestand IBM also now has the 3-HOP DR setup which is even more involved. 
I know IBM and believe HDS plays here. EMC may play but I am told only one 
way. NOTE: everyone looks at mirroring from your site to the DR site but few 
remember about after a DR, how do you get things back to your rebuilt 
location. Thus one would mirror from the DR site back to your host site. 
Ensure any DASD vendor you pick supports it. 

         3) For either 1&2 this is the most expensive solution and a good chunk 
of the cost will be in TELCO charges for OC3's, OC12's, Dark Fiber, etc, 
redundancy, managed services, etc. 

     b. If you can tolerate a small data loss, then ASYNCHRONOUS mirroring 
can be selected, at a reduced cost. This is where IBM , HDS, and EMC all 
play. Distance is not a factor. 

One successful strategy to reduce cost significantly is to lease space in a DR 
provider's facility such as IBM BCRS, Sunguard, etc, and put your DASD on 
their floor and mirror to it. In the event of a disaster, you would just hook 
up 
to a portion of their bigga z9 or z10 and they give you "X" amount of MSUs
(MIPS) and you now run your work. This is by no means an all inclusive 
explanation but should get you thinking about what kinds of management 
decisions they may have. Then once you cost out each desire, it may change 
to a more realistic $$$$ number. Hey, if money is no objective, then go for it 
with some expensive consultants and just start writing checks. Here spending 
your (and my) tax dollars, we had to get the most value from the least 
amount of money. If you want to chat in general, contact me offlist. 

jim 



   

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to