1.004: are you wasting money on your mainframe dr solution?
You are if you've invested in IBM's Geographically Dispersed Parallel Sysplex (GDPS®).
It is widely understood that IBM derives a significant percentage of its profits off of its installed base of "Z" mainframe customers. When there are really no other options for either the hardware or the system software, it is a given that customers will be paying more than if there were real competition.
If you're an IBM mainframe customer, you are - by definition - IBM's profit stream. And there are enough of you that it adds up to some pretty darn interesting numbers.
That IBM has nearly perfected the art of vendor lock-in is inarguable - Mr. Pennybags himself (TonyP) took two whole blog posts to reiterate the ways IBM can legally take away any hope of choice so as to maximize your contributions to their bottom line (see: The Pot and the Kettle and More exploration into Vendor Lock-in - combined they are a comprehensive tutorial on the subject, from an obvious expert).
So, if you need a mainframe to run your business, today you really have no choice but to buy that mainframe (and the operating software) from IBM. For a fair -if not competitive - price, I'm sure (NOT!).
And if you need comprehensive multi-site disaster recovery? Lucky you - IBM will happily sell you a comprehensive GDPS engagement to automate your DR fail-over (and hopefully fail-back). Imagine, months, if not years, of IBM GS folks on-site, tinkering in your environment, crafting and testing proprietary DR automation scripts for you - right before your very eyes. And for no more than couple of million smackers to start (depending upon how much you need to fail over, in both MIPS and GBs), plus ongoing maintenance and consulting fees for upgrades.
Such a Deal!
And you'd better enjoy it, because you really don't have any choice in the matter.
Until now, that is...
"the future ain't what it used to be"
Today EMC announced its new Geographically Dispersed Disaster Restart (EMC GDDR), a comprehensive alternative to IBM's GDPS that costs about a quarter of GDPS to implement and takes a fraction of the time to get you up, running, and self-sufficient!
EMC GDDR offers mainframe customers their first real choice for automated two-site or three-site disaster recovery, leveraging the extensive operational and recovery efficiencies of SRDF, the worlds most widely-installed array-based replication platform.
Unlike GDPS, which is primarily a costly services engagement, GDDR is actually a product - a software solution that simplifies and automates the process of transparently and efficiently relocating entire mainframe workloads to a remote site. With GDDR you can easily automate operations fail-over for planned outages, such as maintenance operations or site upgrades, or you can set-it and forget-it to handle unplanned site outages.
Alternatively, if you don't need all the automation, you can simply choose to deploy EMC's Autoswap and Enterprise Consistency products for a slightly more manual fail-over with full consistency across your mainframe and open systems applications. Both have been available for years, and many customers have already taken advantage of their value in their production environments. Meanwhile, many IBM customers have been forced into licensing full-blown GDPS in order to get even this basic level of automated mainframe workload fail-over (aka HyperSwap) and/or multi-application consistency.
As much as 70% of the GDPS installed base probably only needed what IBM now offers as "GDPS Lite" (and if you're among that 70%, you're probably negotiating a steeper discount on your next purchases from IBM - or at least, you should be). Guess what - EMC's alternative to GDPS Lite is still less expensive...in fact, no matter which way you go,
using EMC's DR solutions instead of IBM GDPS will inarguably
"save you cash - which is just as good as money!"
"you can observe a lot by watching"
If you're an IBM mainframe customer, with or without GDPS, here a few observations that you might find interesting:
- The typical IBM GDPS installation is a 2-site configuration (EMC GDDR supports both 2-site and 3-site - for continued protection upon site failure).
- Estimates are that IBM have deployed about 364 GDPS licenses world-wide since 1995. That works out to about 28 sites, or 14 implementations (at 2 sites per implementation).
- A small implementation might go for something like $2M for the initial licenses and implementation, which would mean IBM is clearing a cool $28M annually. And that's not counting ongoing maintenance and IBMGS fees for upgrades or configuration changes.
- GDDR is targeted to provide the same capabilities for about 25% of IBM's GDPS license AND maintenance fees.
If you ask me, $2M is a lot of sheckles for DR automation - especially if it's only for a relatively small amount of data. I've seen estimates that GDPS averages little more than 2TB per site. But then, if your business is expected to continue 24x7x365 no matter what the disaster and you've grown accustomed to paying the sole-provider premium for your mainframe infrastructure, I can see why you might think it's reasonable.
But you shouldn't have to think like that any more - EMC has brought you choice.
Just like you don't have to be locked in to expensive Fibre Channel drives for 100% of your mainframe storage - with EMC, you can deploy Enterprise Flash Drives for the ultimate in performance and (or) massive 1TB SATA drives for the lowest $/GB for your less frequently accessed data. You can even get Dynamic Cache Partitioning that supports more than two partitions to insulate different mainframe partitions, hosts and workloads from each other.
All shipping today.
"it ain't the heat; it's the humility"
Yogi Berra once said that "I didn't really say everything I said" - but I'm pretty sure he did. At least, he deserves credit for the quotes used in this post (I snagged them from here and there).
One of my favorites is this one:
"If the fans don't want to come out to the ballpark,
nobody's going to stop them!"
Now that you have a real, honest-to-goodness choice in disaster restart solutions for your mainframe environment, I encourage you to vote with your feet and stay away from the Big Blue Ballpark.
In fact, pick any one of the following four options:
- Buy EMC GDDR instead of even considering the over-priced complexity of IBM GDPS.
- Get a bid from IBM for their GDPS Lite (HyperSwap and cross-control unit consistency) – it might be all you need, and it’ll be boatloads cheaper than GDPS
- Better still, buy EMC Autoswap and enterprise consistency – same as #2, but even less expensive than IBM's "cost saving" offering
- If you really can’t stay away from the BBB, then at least force IBM to sell you full-blown GDPS at 25% of whatever they quote on their "final offer"
You've been funding IBM's bottom line for way too long - it's time to start biting the hand that's been wrapped around your wallet for all these years.
Oh, and to IBM and it's well-worn track record of vendor lock-in, I say:
"I want to thank you for making this day necessary"
technorati tags: IBM,GDPS,EMC,GDDR,Symmetrix,disaster recovery,disaster restart,hyperswap,remote replication,SRDF,zSeries,IBM z,Series z,consistency,congroups,autoswap,vendor lock-in,z/OS,mainframe,disk,storage,arrays,IBMGS,global services,Managed Storage Services,TonyP,Monopoly,Mr. Pennybags,sole-source,proprietary interface,enterprise storage,multi-platform DR,Yogi Berra
Comments