Tuesday, March 17, 2015

Can you believe it? “Pragmatism trumps spectacle”!

Influenced by the recent post of Margo Holen, the Internet of Things certainly is attracting more than its fair share of the media spotlight. However, it’s also fair to say that with more and more things communicating, your primary systems just can’t fail!


For true car racing enthusiasts, including those who still harbor hopes of becoming a world champion, Racer is the must-read magazine each month. It provides not just in-depth analysis but pages of data about team and driver performances. And oh yes, the pictures look really cool, too. Margo and I are about to start our eighth year of High-Performance Driver Education (HPDE) track outings across the western states of America and, while using the word race remains taboo, we are on track at least four times a day for 20 – 25 minute “sessions”. Having said this, we are amateur enthusiasts at best and only on a good day.

The picture above comes from the very first post of June 6, 2008, to our social blog, Buckle-Up-Travel, Off to the races ... and it continues to surprise many that it is Margo behind the wheel of our C6 Corvette; a vehicle that has been in production for more than sixty years and has come to define what today we call America’s sports car. Muscular, flashy and always ready to put on a show! But as I read the latest issue of Racer magazine I came across a brief editorial page with the heading, “Pragmatism trumps Spectacle” and my thoughts turned to the more than forty years that Tandem computers, now NonStop systems, have been supporting the most mission-critical of all business applications.

NonStop systems have stuck steadfastly to the task of providing 24 X 7 NonStop availability, with near linear scalability, but without the flash or even the fanfare that today heralds the arrival of something new.  Announcing the Apple Watch, as Apple has just done with its high-energy, spectacular, Spring Forward event immediately after the switch to daylight savings (quite the marketing coupe that grabbed everyone’s attention) was a flashy affair but shortly HP will be officially taking the wraps of the next family of NonStop systems – NonStop X. I doubt there will be much to show by way of fanfare and bright lights, but the message will ring true. If you have mission critical applications to support, where failure is not an option, then NonStop delivers.

The NonStop X family of systems is arriving at an opportune time – much faster performance than many of us had thought back in 2013 when the wraps were taken off the program for NonStop to embrace the Intel x86 architecture . But then again, occupying center stage of many companies transaction processing, it’s going to get really busy really soon and faster systems will prove a blessing. No more so than when it comes to the changing landscape of mobile device usage. In his article “Survival of the fittest – NonStop in mobile networks” published in the May/June 2014 issue of The Connection, HP Telecom Business Development Manager, Rene Champagne, observed how “Modern usage of mobile services has shifted from voice and text as the primary services, towards mobile broadband services. Real-time services based on smartphone apps are the defining characteristics of this decade.”

It’s hard to argue against this observation and the importance of this shift, Champagne went on to note was that, “The real-time nature of these apps makes service interruptions immediately apparent as social media updates cease to operate, and as services such as cloud storage become unavailable.” When it came to quoting the results from studies undertaken by the European Union Agency for Network and Information Security (ENISA), specifically their second Annual Incidents Reports, Champagne highlighted that availability was still very much an issue. “The determination that the root cause for most incidents was ‘System failures’, and that ‘hardware failures’ were the most common root causes,” Champagne quoted from the report even as the report  noted how this determination, “diverges from general IT norms which generally find that operations and administration, and software failures are the most common outage sources.”

Remaining steadfastly focused on the task of providing NonStop availability with near linear scalability for forty plus years may not have been a bad thing to do, after all. Even as system failures and hardware failures happen less frequently, the impact of failures observed by many customers can quickly become a different kind of spectacle, and one quite likely to make the evening news. The increased dependency on mobile devices and their ability to initiate mission critical transactions is only the beginning as it will not just be humans on the other end of the line but machines.

The impact of the Internet of Things (IoT), and in particular, as HP Master Technologist, Justin Simonds,  has been highlighting in his recent presentations, the Internet of Mission-Critical Things (IoMCT), where outages not only will disrupt production lines and transportation but healthcare and even security will be hard to ignore. One thing we can be very sure about is that with the arrival of such streams of information, changes will be constant and necessary. As we transition to the world of small, highly focused, apps keeping up with it all will be a challenge for all but the most mature of infrastructures.

In the real-time IT world systems, platforms, operating systems, middleware and applications are all providing updates about their operational status, I suggested in the post to the WebAction blog, In a Realtime World, NonStop Customer Experience Can Change in an Instant! A constant barrage of data makes tracking the performance of an application difficult; who can tell whether basic SLA metrics are being met? Add into this the added complexity that comes with hybrid computers, not to mention the use of clouds and the task befalling those responsible for ensuring customers are being served in a timely manner appears more like black magic than pure science. Yes, availability still matters and HP being pragmatic about this key attribute of NonStop is something we can all be very proud of – outages today are simply unacceptable.

The support of things, the majority of which are likely to be machine sensors and instruments including the many such devices present in our cars, may not all be supercritical or warrant the supporting technology infrastructure that we associate with NonStop, but with increasing legislation and a healthy appetite from some market segments such as insurance, one man’s nice-to-have may very likely be another man’s must-have and command the presence of the most available systems possible. In an earlier post to the WebAction blog, How Will the Internet of Things Impact Streaming Data? WebAction Director of Marketing, Jonathan Geraci, made similar observations. Referencing research by Forbes journalist, Mike Kavis, Geraci states that, “This sentiment is aligned with our thoughts that we are moving away from a software world of ‘query/response’ (ask a question, get an answer) and moving toward the sensing enterprise.”

Building on this sentiment, Geraci then goes on to suggest that, “In the sensing enterprise software monitors your streams of data and acts autonomously to specific correlated events in your streams. The IoT drastic increase in the number of connected devices will dramatically increase streaming data volume, velocity, and variety. The next logical question is how will you handle all of the streaming data from the Internet of Things? The WebAction platform provides stream analytics applications that are extremely efficient and easy to use.” And yes, as we are now seeing, WebAction supports the many logs and event files that are maintained on a typical NonStop system.  

In my most recent exchange with HP’s Simonds, he reiterated the value that comes from being as pragmatic as the HP NonStop R&D team has demonstrated to be through the years. When it comes to NonStop systems, everyone involved knows of the “very clear mission – to build a computer that won’t fail.  All our development is in sync with the mission and every developer hardware and software has to answer the question ‘what happens when this fails?’ Not if, but when.” Yes, all those years ago, Simonds notes, “Tandem was the first to have no single point of failure.  That has been copied.  Tandem was the first to have online repair.  That has been copied.  But NonStop after 40 years is still rated by IDC as an Availability Level 4 (AL4), a full level higher than any clustered system.  NonStop still has the lead but many systems are catching up.  That’s why I am very happy that Wendy is relooking at Indestructible, Scalable Computing.”

When Margo and I first turned a wheel on a race track it was in a fully automatic Corvette – the C6 Corvette in the picture above - it was for very pragmatic reasons; it allowed us to concentrate on learning the track. Trust me; when it’s your first time on a track, you are looking everywhere to make sure you don’t hit another driver. With the volume, velocity, and variety of data streaming into our data centers, having a system rated an AL4 lets you worry about the processing and all that such processing entails without having to look down at the console to see which process is actually running and whether failures have occurred at a critical time. Can you believe it? NonStop thriving after all these years?

To be truthful, perhaps we should be going back to our CIOs and asking how alternate offerings survived at all – being pragmatic, of course, perhaps CIOs already know of this and have a lot more to do with the longevity of NonStop than we care to acknowledge. Maybe NonStop isn’t as big a secret as some of us suspect and maybe, just maybe, a spectacle isn’t really what a CIO is looking for today! 

No comments: