Clouds go offline; data centers are compromised; critical infrastructure is held to ransom. Thank goodness for today’s modern HPE NonStop systems that we can run any which way we like!
“Crash your car?”
“What?”
“Would you like me to crash your car?”
You can imagine the look on my face as this dialogue
continued. True, my hearing may not be as good as it once was and true, the
cars I drive are low and working any sort of credit card reader through a car’s
window can be problematic but seriously, crash my car? As it so happened to be
on this occasion the attendant at the car wash was only asking me whether he
could “cash my card?” He happily took my credit card, inserted it into the
machine and we were on our way.
For any member of the NonStop community, the above
exchange might arouse a wry smile or two but it continues to be crashes of
another type that only create a shaking of the head. In this day and age,
surely systems shouldn’t be crashing for any reason especially those we
interact with on a regular basis. Credit and debit cards should be easy enough
to process as should any perimeter device we interact with to complete a
financial transaction. You would think that after four decades the concept of
fault tolerance might have remained a priority for every business.
The sign on the window said it all. “Unfortunately we have had a system crash and we are unable to open for business at this time.” The store was in Ft Collins and catered to those looking for anything to do with their kitchen – appliances, storage containers, condiments, jams and even homemade chocolates. It was all available at least when they weren’t experiencing a system crash.
Margo and I often visit this store with no
particular goal in mind and just as often we come out with one purchase or
another – it’s the only place where I can get legitimate Coleman’s Mustard
that’s the equivalent in strength to what I relied on in Australia. Yes, I
often get the powder and mix my own mustard to ensure the heat is elevated to
the correct level for mustard.
On this occasion, our trip into Ft Collins was to no avail. And here’s the thing – not only did this store loose business but it disappointed loyal customers. I returned to my car wondering, how many other possible patrons they disappointed or worse have discouraged from returning.
After all, there was Boulder’s famous “Peppercorn” establishment right on the
Pearl Street Mall that provided the same products as did Ft Collins “The
Cupboard.” Fault tolerance is not a
religion or a philosophy it’s what HPE NonStop systems deliver and in a world
that is increasingly being enticed to enjoy the cloud experience, have we
become blasé about crashes?
What we have seen rising in occurrences is applications
that crash. In an April 9, 2021 post to the blog, Decode, Mario
Žderić wrote of how, “It’s almost impossible to keep
any customer satisfied if your app crashes regularly … Such experiences during
financial transactions may also impact the buyer’s trust in your brand
significantly.” Žderić then describes 11 reasons for applications to crash some
of which apply to even the most robust of enterprise applications. Probably
more so when those enterprise applications happen to be mission critical.
Among the reasons for applications to crash – Žderić
writes mostly about mobile phone apps but the reasons for some crashes applies
equally to enterprise applications:
Inadequate testing – “It’s impossible to deliver your
users a stable and reliable app without thorough testing. To prevent or
minimize your app’s risk of crashing, it’s important to test it continuously.”
Really? Still doing it by hand, surely not!
Software development life cycle – “At each step of the
way, there are stage-specific issues that could compromise the stability of
your app.” Stages? Yes, from implementation through testing, maintenance and
then subsequent planning for what comes next, gaps can open up and critical
processes be missed!
Errors Due to Agile Methodology – “Agile’s underlying architecture focuses on delivering new product features and updates in small increments at regular intervals … this also comes with its disadvantages from a technical perspective: With incremental deliveries, you often have to deal with compatibility issues, poor resource planning, inadequate documentation, and indefinite improvement rounds.”
Disadvantages? This is not a discouragement of
Agile but a reminder that Agile is no excuse for poor management and indeed lax
or overambitious perhaps borderline aggressive oversight!
Fast forward to today and to the cloud experience and the reasons for crashes have become far more sinister. Everything from a simple mischievous hack to a far more concerning orchestrated deployment of a virus associated with ransomware: Could NonStop prevent such intrusions?
The reality
is that it can – as more and more legislations are passed by governments
concerning privacy and in truth, protection of sovereignty, cloud experiences
may cloud the real issue. For the NonStop user, there are plenty of processes
that may indeed be easier and more effective to run in clouds – backup (and
restore), management and monitoring tools and utilities, analytics – but the
applications themselves?
There is a reason why NonStop transcends all other
approaches to running mission critical applications. The very architecture that
underpins the fault tolerant nature of NonStop in and of itself ensures crashes
do not occur. There is a message system and above it a file system that
together ensure rogue code doesn’t suddenly appear nor can data suddenly be
impacted by outside sources. You can put ropes around your NonStop system but
those optics don’t tell the full story. There are ropes for sure but they are
virtual, a function of the message and file systems and as such remain out of
sight from the end user.
Perhaps what is also underestimated is the value provided by the integrated software stack that makes up a key portion of what NonStop is all about – there is a lot more to fault tolerance than just stringing together redundant processes and duplicated links. Applications built upon the NonStop middleware, be they written in TAL, C/C++, Perl, Java, and much more, are given an assurance of working once tested.
The user can readily
address deterministic bugs but NonStop easily sorts out those particularly
tricky to repeat non-deterministic bugs. This was all fully disclosed back in
the days when NonStop championed the NonStop Availability (NSA) initiative.
Systems continue to crash. Clouds can evaporate and increasingly, restrict access to mission critical data that lies beyond the reach of the user. Apps can become unreliable even as networks struggle to provide coverage as well as bandwidth. The point is that there are no absolutes with technology.
There are no guarantees. NonStop can do so much to minimize
much that ails today’s cloud computing; NonStop can run happily within a
private cloud and when hybrid clouds become a consideration, there is a future
coming where NonStop will treat the presence of clouds as they do virtual
machines – cloud 0, cloud 1, etc. And why not?
If you want to make your cloud experience enjoyable, reliable and indeed, crash resistant there are already ways that you might want to consider to do just this and NonStop should be part of that discussion. The NonStop team is always responding to changing enterprise needs and when it comes to ways to eliminate downtime on traditional, virtual or hybrid systems, come up with a good reason for the need and they will listen.
Just as they have
been doing for four plus decades? Just ask yourself; is today’s NonStop resembling
the NonStop of the 1970s/1980s? No, nothing like it and that’s because the
changing enterprise requirements of IT have changed and NonStop has
demonstrated an ability to evolve to meet those needs.
I didn’t give my car to someone who asked to crash it even as it was a reminder that I might need to have my hearing checked. On the other hand, when I was informed of a system that crashed, I couldn’t buy my favorite mustard.
Crashes may not be a part of the NonStop vocabulary even as we are cognizant of what can cause the crash of an application. But the game is changing and the changes have become games ranging from mischievous to outright villainous.
There’s no better time to include NonStop in the conversations even
as there’s no better time to interact with your peers. Perhaps there will be
answers provided at the upcoming NonStop Technical Boot Camp (TBC) – will we
see you there? Most important of all however is the need for more NonStop
champions as we have the experience; we have the knowledge and yes, we still
have that level of enthusiasm that warrants others to dig deeper into what
NonStop can do for them today.
Comments