My primary concerns around the idea of "the 9's". It is undeniable that
systems built with sufficient redundancy can achieve 99.999%
availability without a single point of failure on a single system. z/OS
is renowned for its fault tolerance in both the hardware and software
realms. Nonetheless, it's not the sole platform capable of providing
99.999% or higher availability with the aid of clustering. Here is an
example:
https://buy.hpe.com/au/en/compute/mission-critical-x86-servers/integrity-superdome-x-servers/integrity-superdome-x-servers/hpe-integrity-superdome-x-server/p/7161269
But what happens when we encounter outages?
Scheduled Outages: These include maintenance or weekly or monthly batch
processes that require the shutting down of online systems. If you
haven't invested heavily in clustering a high availability system stack,
outages will be inevitable.
Unscheduled Outages: Errors in the system! System software or
applications may crash, making them unavailable until they're either
rolled-back or fixed. The fix might take hours. Your 99.999%
availability is compromised if your CICS transactions start crashing or
a crucial batch job damages a database table and necessitates recovery.
Of course, such instances never occur in a utopian world, right? Anyone
with mainframe experience knows about IBM's red alerts
https://www.ibm.com/support/pages/red-alerts-subscription-service-zseries-0.
Human Mistakes: This is a favorite of mine :)
Take the case of Air New Zealand - Certainly, airlines require 99.999%
24/7 systems. Just don't entrust IBM global services with your mainframe
management: https://www.theregister.com/2009/10/12/ibm_new_zealand/.
The NatWest/RBS CA-7 debacle - A rookie system programmer failed to
notice the unsuccessful upgrade of CA-7, which led to a three-day
outage. This resulted in a hefty fine for NatWest by the UK government
https://www.theregister.com/2012/06/26/rbs_natwest_ca_technologies_outsourcing/.
So, how's that 99.999% looking?
The NAB outage caused by a power failure: I was once an NAB customer.
The so-called 99.999% reliability was a myth! During my time at SHARE, I
couldn't withdraw money from an ATM due to their weekly maintenance
https://www.itnews.com.au/news/nab-outage-caused-by-power-cut-to-mainframe-491780.
Further issues with NAB can be found here:
https://delimiter.com.au/2010/12/09/what-the-hell-is-going-on-at-nab/
After experiencing several disastrous outages, NAB has decided to blame
these interruptions on outsourcing (IBM GS/Kyndryl) and seeks to
transition to AWS:
https://www.afr.com/technology/nab-raises-the-white-flag-on-it-outsourcing-20190508-p51l6k.
Hopefully, this will not be a quick transition as they are one of customers!
A mainframe failure caused a six-hour HSBC customer service outage, with
reports of stolen account information for approximately 15,000 active
clients at the bank's Geneva branch. Consequently, HSBC must have had a
9.99% availability that year and the no mainframe data breachs myth is
totally busted:
https://www.datacenterdynamics.com/en/news/mainframe-outage-brings-hsbc-customer-services-down/
And these are just a few examples. Mainframes are insanely reliable but
humans are not.
On 23/5/2023 11:40 pm, Tom Brennan wrote:
"It’s always a good idea to check who wrote the piece and whether they
have a bias. (Or financial interest)" Bill Johnson - 5/19/2023
On 5/23/2023 8:19 AM, Bill Johnson wrote:
I remember a while ago I mentioned the mainframe guaranteed 5 9’s of
availability and some here mocked me. 99.999 had been true for quite
some time. Now it’s 8 9’s. (99.999999) of uptime per year. No other
platform comes close. Like I’ve said dozens of times, the mainframe
is growing & will be the platform of choice for decades into the
future. Regardless of how crowded the parking lot in Poughkeepsie is.
Sent from Yahoo Mail for iPhone
On Tuesday, May 23, 2023, 10:57 AM, Glenn Wilcock
<wilc...@us.ibm.com> wrote:
This post from Ross is a must read for this topic:
https://www.linkedin.com/pulse/results-mainframe-application-modernization-migration-ross-mauri/
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
