Jay Howard wrote: >Being the original poster, we currently do partner with a company for our
>DR needs. We are looking for ways that we can reduce our DR costs and >FLEX-ES is one of the options that we are investing. I think we've got two separate concepts running through this thread, so here's an attempt to clarify. There are DR companies -- Sungard and IBM BRS are good examples -- that sell DR contracts of various kinds. I'm going to disagree with the implication that these contracts are "worthless" because of "first in" policies. There were several disaster declarations during Hurricane Katrina, and the DR vendors seemed to do pretty well by all accounts. The contracts have various SLAs and prices -- basically "first" doesn't mean "forever" -- and the (good) DR vendors build enough infrastructure and test regularly with their clients in order to handle even wide area disasters. There's also the separate idea of finding a like minded company in order to strike a one-to-one private partnership along similar lines. In the era of Capacity On Demand that's quite reasonable for mainframes, so I don't really share the concern about lack of production capacity at your partner's site with a disaster declaration. (All bets are off for other platforms.) If the disaster affects both companies simultaneously then that's bad, so you'll probably want to pick a DR partner that's "close but not too close." In many cases striking a private partnership is not a direct cost item. (You still have the costs associated with preparation and testing, but that's always true unless you plan on not having any DR strategy.) So I'm not sure what you mean by "reducing our DR costs," because I'm thinking "zero" here. It's a straight up trade: I'll let you use our production system (spin up an LPAR/COD) if you have a disaster if you'll agree to the same for me. I'm sure there's a reasonably comparable mainframe shop elsewhere in Georgia. Many companies opt for multiple DR arrangements concurrently. Some simply need their own GDPS. I know of an insurance company that understands it would have to declare bankruptcy if there was a system outage lasting more than 60 minutes. There are probably several companies with even less tolerance for outage. Either of these arrangements (DR vendor or private DR partner) is better than nothing. Better than nothing may be sufficient. Nothing is what an awful lot of companies have right now, and (prediction) some of them will effectively go out of business when disaster strikes. Do note that controlling DR costs is what mainframes do exceptionally well, so cost context is important here. Building and testing DR infrastructure with other platforms can be brutally expensive, and what you end up with isn't as capable anyway. A big part of the reason is Capacity On Demand, but it's not the only reason. In fact, I can imagine many cases (Linux and J2EE especially) in which mainframes should act as the DR systems even if the primary production systems are not mainframes. - - - - - Timothy F. Sipples Consulting Enterprise Software Architect, z9/zSeries IBM Japan, Ltd. E-Mail: [EMAIL PROTECTED] ---------------------------------------------------------------------- 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