There’s an air of excitement surrounding Apple that I haven’t seen around technology since Sony introduced the Walkman in the late-1970s and Motorola introduced the original Razr in the mid-2000s. And the picture above is of the wonderful display from flowering snowberry trees that surround the walkway to our house and yes, provide a fitting background to another technology marvel – the legendary Nissan GT-R, or Godzilla – that is now Margo’s daily drive.
The iPod shook up Sony, but with the iPhone and then iPad Apple is shaking up the whole industry – redefining the PC and refocusing the world’s attention onto smartphones and tablets. As another blogger expressed it recently, “looking backwards at the iPad disruptive innovation did happen. The iPad was a disruptive innovation. It was not predicted beforehand. It created a new product category.”
However, it is the ease of access to iCloud that has really captivated the marketplace and for those happy with the controls Apple has in place, being able to move from phone to tablet to laptop to desktop and enjoy the same user experience, picking up tasks exactly where they were left off – no matter the Apple device being used at the time – is for me another example of disruptive innovation.
Discussions about clouds have been going on for some time with Amazon.com and Salesforce.com, but almost overnight Apple has been singularly successful in moving discussions about clouds into the bigger, consumer, marketplace such that the value proposition has become perhaps even better understood.
HP is not blind to what’s going on either – mobility and cloud computing are at the top of the priority list driving their roadmaps. From product to services to consulting, pretty much every division within HP has a message in place involving clouds. The NonStop community is not being left behind or being ignored – quite to the contrary, there’s some serious heavy-lifting under way to ensure a role to play for NonStop in support of cloud computing.
Leafing through the pages of the March – April issue of The Connection, I came across the feature “Persistent Cloud Computing (PCS) Architecture” by Master Technologist for Enterprise Solutions & Architecture (ESA), HP, Justin Simonds. “Cloud computing is one of the hottest topics in IT today. Articles abound on how cloud computing is a potentially disruptive architecture and how business will or will not, take advantage of this new computing paradigm,” Simonds open the feature, and he then adds “notorious outages and failures have recently occurred within public cloud systems have sparked huge interest in how and where HP NonStop can help customers in their journey to cloud services.”
In his post of October 2nd, 2011, “The ‘Big Five’ IT Trends of the next half decade: Mobile, social, cloud, consumerization, and big data”, that appeared on the web site of publisher zdnet.com, blogger, Dion Hinchcliffe, opened with “of all the technology trends on this list, cloud computing is one of the more interesting … among the large enterprise CTO and CIOs I speak with, cloud computing is being adopted steadily for non-mission critical applications and some are now even beginning to downsize their data centers.”
Perhaps of more interest to the NonStop community were Hinchcliffe’s closing comments on cloud computing, when he noted that “until cloud computing workloads can be seamlessly transferred back and forth between a company’s private cloud and public / hybrid cloud, adoption will be held back and favored largely for greenfield development.”
Really? With this in mind I went back to Justin Simonds for more insight about the feature he wrote for The Connection. “Cloud computing is ‘the’ topic today. About two years ago, Tom Miller, Keith Moore and I were working on a project where we wanted to bolster x86 architecture with some of the NonStop fundamentals and the idea of extending Pathway through an API came to mind; we all felt that NonStop involved in a hybrid architecture approach was a good thing,” Simonds explained.
“Wouldn’t it be cool to have NonStop overseeing a cloud environment? Since we were pushing servers out to a Linux server, which could be physical or virtual, it did not seem like a stretch to have the serverclass located in a cloud. Once we had that working, the same stuff we were doing with the local Linux servers could be done in the cloud. That is load balancing, failover and elasticity,” Simonds then added. “We demoed this at Discover in 2011 and everyone seemed fascinated.”
A subsequent exchange with Keith Moore provided even more insight into the program as, according to Moore, “our general vision all along is that most current cloud server instances support discrete application execution instances. We are thinking that this use of “the cloud” is somewhat limiting. Most of the current enterprise cloud deployments use application templates for creation of specific vertical enterprise functions. To the extent that existing enterprise applications currently work together in an integrated way, these are still deployed as mapped instances onto virtual cloud instance.”
But then, returning closer to home for the NonStop community, Moore suggested how “this is tremendous opportunity for HP and others to address a need for managing and monitoring these specific user needs. The benefits of moving to this sort of ‘templated-cloud’ are fairly well documented and often represented in a leveraged, cost-lowering business driver. However, if you look at legacy and future application development and deployment models you see that for most large enterprises, there are issues with inter-application complexity that cannot be simply resolved by deploying templates of vertical business applications.”
Reflecting on what Simonds too had suggested, Moore then continued on this theme as he pointed out how “the workloads in most pan-enterprise computer systems need to support dynamic, self-healing business functions. By using NonStop as a service manager for the cloud, there is significant value to almost all enterprise application scenarios presented for cloud usage. NonStop adds fault-tolerance, dynamic scale-up/down, and security/audit capability to real-time activities that might use cloud server instances of business applications.”
Digging even deeper into the topic, Moore then hit on what I believe could propel NonStop more visibly into the discussions taking place between CIOs and CTOs when he proposed how “in many enterprise applications, the usage is so dynamic, variable, and vulnerable when on the cloud that it needs some sort of external ‘daemon’ to manage and monitor the cloud instances. NonStop, using the fundamental capabilities of TS/MP to manage response time and availability, delivers this ‘cloud daemon’ capability. NonStop PCS uses standard NonStop TS/MP to deliver an always-on daemon for business application services.”
Much was written about NonStop and cloud computing following the demonstrations featuring NonStop interacting with clouds, private and public, at last year’s HP Discover event. My post on the subject, “NonStop revels in Clouds!” attracted more hits on my site than any previous post, and to me this was an indication of how topical clouds had become even within the NonStop community. Not a product and not strictly a service, Justin, Keith and Tom then provided education and put on workshops and more recently, began their initial pursuit of Proof of Concepts (PoC’s) that have already attracted considerable interest.
“We just completed a PoC for a financial institution, where Base24 was used to drive Java servers using PCS. The configuration consisted of two NonStop blade servers and a Proliant RedHat Linux server, all running the same Java application,” explained Tom Miller. “In all of the tests the application employed the JDBC driver connected to the database of record, which for this customer, was an SQL/MX database. We tested two configuration policies; load sharing - where all three nodes evenly shared the BASE24 client traffic, and a capacity burst scenario - where, as the primary NonStop node reached a resource threshold, TS/MP 2.4 began distributing client traffic to the other two nodes.”
“The customer was extremely satisfied with the results. We were able to achieve > 1200 TPS with both configurations with an average response time of 10 ms ~ 2ms,” Miller reported. And for me, this proved to be a reality check. Yes, the early observations about getting NonStop involved in this way and to have it a part of a hybrid architecture, were beginning to be realized and looked to be paving the way for possible future consideration by others within the NonStop community.
As I was wrapping up this post I went back again through the comments Simonds and Moore had provided to me and what was hard to miss was their enthusiasm over the potential benefits NonStop could provide. “I like to tell folks that if they understand the difference between deploying a business ‘server’ and a business ‘service’, then you can understand the need and the capability of PCS on NonStop,” Moore had said.
“NonStop is, and always be all about providing a business SERVICE to enterprise clients. PCS is envisioned as a tool to deliver a NonStop-level of SERVICE to more than just NonStop servers.” Simonds then closed by saying “writing the article was a way to get the message out to a broader audience and to let everyone know NonStop plays, or should I say revels in the clouds, especially with all the HP cloud announcements going on.”
And in the coming months, I anticipate hearing even more about clouds as we move deeper into the summer months and it will be with a lot of interest that I walk the exhibition floor at the upcoming HP Discover event – what more surprises are in stall from the team of Simonds, Miller and Moore!
Comments
Justin