Last weekend, driving west along California’s highway 101 I passed the local Mercedes Benz dealership, and a quick glance at their showroom had me heading for the exit immediately. A quick U-turn and I was in their parking lot.
The picture above captures the sight – a new Mercedes SLS “Gullwing” sports car! Alongside the SLS, a car I hadn’t seen while driving on the highway, was a Mercedes SLR “McLaren”! Both cars obvious works of art, but beneath the metal and carbon fiber skin a full complement of race-bred technology. The picture above captures the two cars, side by side, with their distinctive doors fully extended.
For years, I have viewed this marque as the benchmark when it comes to quality engineering in an enjoyable, reliable package. It was in the early ‘80s that, while I was living in Sydney, my neighbor gave me the keys to his SL350 for the day, and with the coupe’s retractable rag-top pulled back I had a blast cruising the local beachfront streets. I even used it to make a business call at a data processing service bureau I had just landed as a client.
I have never owned a Mercedes Benz, having always preferred the cars produced by their competitor in Munich, but I always valued the information that came from driving their cars. Through the years the list of features has grown enormous as has the power they have been able to generate from their engines and yet, as a benchmark, I am no longer as sure as I once was that the benefits of owning a Mercedes Benz is worth the price they charge.
It’s been hard to escape the many stories circulating of how, in the March timeframe, there will be major announcements by Leo Apotheker, HP’s newly-appointed CEO, on everything from the make-up of the board, his executive team, the products, as well as a new strategy for the company. While there has been considerable speculation on what will be addressed, we have already seen some of the details as the changes in the makeup of the board of directors became public.
When it comes to strategy and to technology and products, for many within the NonStop community, it’s as if the news being leaked has been telegraphed for some time. With the push to commodity hardware almost complete, and where very little of the hardware package that makes up today’s blades required to run NonStop is unique, it will surprise none of us should the hardware package become ubiquitous. After all, it is the full stack that is NonStop where the magic has always existed.
Along these lines, in the feature “NonStop, A Running Commentary” just published in the January 2011 issue of the Tandemworld.Net e/Newsletter, I even went so far as to suggest that, with the way the NonStop Enterprise Division (NED) R&D has handled the adjustments that had to happen with the switch of chip vendors to Intel, I have to believe NonStop could run on a pair of sardine tins, if there were any processors inside whatsoever!
Nostalgia? It’s not the hardware anymore that differentiates today’s NonStop, so let’s not get hung up on this issue, I suggested!
For the many of us relying on NonStop servers to anchor support for mission critical applications we have been aware that NonStop is about a lot more than just the hardware, and even more than just the operating system. If it had been just about the operating system, then the history of NonStop would be long over, and the 35th anniversary the stuff of imagination. Perhaps outlasting some of the better-known operating systems of the late ‘80s, all the same, but eventually losing relevance in the face of power house operating systems such as Unix and Windows.
Just as I have considered the Mercedes Benz as the benchmark for quality engineering in an enjoyable, reliable package, so it is that I view the complete stack on NonStop as the benchmark for continuous, perhaps even permanent, availability. From the transactional TS/MP (Pathway) layer of the stack on up through the Transaction Management Facility (TMF), the NonStop SQL/MX data base and even the Remote Database Facility (RDF), these layers when exploited in combination, ensure modern applications can be deployed in circumstances where downtime for any reason simply is not an option.
In the December 2010 issue of Database Trends and Applications, with its theme of “Big Data, Big Issues” in an article entitled “In Search of the Elastic Database” the author writes about the Relational Database (RDBMS) as the data management layer underpinning every popular database implementation in use for the past two decades. However, in addressing this topic, the author raises a number of concerns, among them, the ability to handle really large volumes of data.
“Unfortunately, unlike most other elements in the application stack, relational databases scale out very poorly. With a few exceptions, it’s not possible to simply cluster your database and add nodes to the cluster as workload increases. Relational databases work best when all the data is managed by a single node.” The author then observes how “shared disk clustered databases such as Oracle’s Real Application Servers (RAC) can increase or decrease cluster members faster, but have an unattractive licensing model.”
Continuing with his observations on Oracle RAC, and looking ahead to possible requirements of cloud computing, the author then remarks on how they “are generally not deployed with very large node counts, and do not solve all the issues involved with scaling I/O rates and data storage. So, purveyors of cloud-based computing infrastructures also need an elastic database.”
NonStop systems aren’t just about the hardware. They aren’t even about the operating system, or about the database management system. NonStop systems are about the collection of layers that in combination make a complete stack that leverages it all, enabling a single node to take on enormous proportions, and for the database to scale well beyond competitive implementations whose origins were in systems that maxed-out with just a few processors.
NS SQL/MX as a database system shows up on very few industry analysts radars. The list of features may even appear less than ideal when lined up against the major vendors’ offerings. However, as part of the complete stack that is NonStop, its ability to be as “elastic” as we need surpasses anything on offer today and continues to be among the best-kept secrets that today is NonStop.
As we head towards March when more information about HP’s strategy is unveiled, it’s no secret how much of the focus of this strategy will center on HP’s transition to a software, services, and integration. Building commodity boxes HP can do blindfolded but that’s not where it’s future lies.
Taking these commodity boxes, building integrated systems on top of them, and offering complete stacks to Java and .Net developers alike, is where HP hopes to excel and build a more profitable and sustainable business. That’s where the value for the user community will lie and with NonStop as part of this strategy, there will be renewed focus on the capabilities of NS SQL/MX much to the delight of a much-maligned cadre of supporters!
While my local Mercedes Benz dealer showcased the SLR with its McLaren heritage and the SLS with its (arguably) Dodge Viper influence, the technologies of these vehicles were priced well beyond the reach of most car buyers. However, alongside these two models, was the more affordable and yet in some respects more modern SL 6.3 coupe. Arguably, the value deal!
When it comes to NonStop showcasing an S-Series, or even an earlier Himalaya K-Series, alongside today’s Blade systems, it’s easy to consider them the benchmarks for reliable, continuously available servers.
However, perhaps not as readily recognized in today's modern Blade server running the full NonStop stack is what hides beneath the covers, like the coupes from Mercedes Benz, invisible to the unknowing observer - a multi-layer stack of integrated, “mission critical” bred technology that remains unchallenged after all these years!