Back on August 20, 2007,
coming off my time as ITUG Chairman, I had no idea that I would be still
writing posts for this blog. But here it is, ten years later, and the
fundamental message of NonStop has changed very little …
Right now it feels as though I have been working 24 x 7 as I finally return to my office after almost ten weeks on the road. Living and working from our company command center through thick and thin, as we experienced outages of unimaginable proportions, it is good to be sitting behind a desk in a room where there is a sense of permanency. This office isn’t going anywhere nor is my desk and my computer. However, to suggest to you that I have been working around the clock, 24 x7, just to complete this post, may hold some truth given this topic of 24 x 7 has been on my mind for a long time, but in truth, there has been a lot of down time. The office may be a permanent feature of the home but I will still come and go as business needs dictate and that will rarely ever be 24 x 7.
The almost blasé dismissal of the value proposition of 24 x 7 was a wake-up moment for me. An epiphany of sorts, actually! Have we arrived at a point in time where business continuity isn’t all that important anymore? Across the NonStop community the attention we give to the value that comes with being able to run our mission critical applications continuously, 24 x 7, ensuring our customers have access to the applications they need to access any time they want is sacrosanct. After all, with all that we know today about NonStop would it still interest us if it didn’t run 24 x 7? In my last post, New home, new office; time to take stock … I extoll the virtues of scalability and as much as I really like the value scalability brings to any discussion on the value proposition of a system, it still plays a role to aid availability for most stakeholders in the NonStop community as you do not need to stop your operation to add capacity!
In being blasé about 24 x 7 this week, the analyst with whom I was working did think about the value proposition 24 x 7 provides, but then added that truthfully, it is not fated as strongly as it used to be, as even now, just as with real time, the actual time implications for both 24 x 7 and real time are similar. A few minutes here, perhaps an hour once in a while, surely doesn’t matter in the grand scheme of things. No down time means just that – no down time. Likewise, continuously available sounds a lot like permanent availability and readers of posts to this blog will have come to see that I have made reference to permanent availability many times in the past. Marketing types don’t like this term as it has a sense of no wiggle-room about it and seriously, how can you be sure nothing will fail? However, that’s the whole point – NonStop users keep purchasing NonStop systems because they provide permanency to where their clients will always be able to interact with them.
Like my office and the sense of permanency it affords me NonStop systems can be relied upon to be there even if my own access to the applications is casual at best. However, what I am pushing back on these days is terms like “near real time,” as well as “highly available is good enough!” The origins of real time were in missile detection and interception and there could be nothing vague about that. If you didn’t respond in time and neutralize the threat, you simply ran out of time. Nothing real time about pressing the launch button too late to protect a determined target! Again, this is all well-known to the NonStop community and yet, if industry analysts are becoming blasé about this key attribute of NonStop systems and the solutions that have been designed to run on NonStop, then others will follow and it is most definitely the beginning of a slippery slope leading to enterprises associating any value whatsoever to continuous availability.
A couple of years back, IDC did a pretty good job defining different availability levels with availability level four (AL4) representing the very pinnacle of availability. The only vendors to have systems awarded AL4 were HPE with NonStop and IBM with special Parallel Sysplex configurations of the mainframe. The inclusion of IBM really was for appearances sake – IDC just didn’t want a level of availability where only one vendor had achieved that goal. On the other hand, simply trying to build a Parallel Sysplex configuration for a reasonable cost that supported AL4 has proved allusive to almost every mainframe installation. Adding more components – more moving parts, of you like – to an architecture that never intended to support operations 24 x 7 is hazardous at best, but more often proves foolish at worst. If AL4 could only be awarded to vendors with systems that provided the level of continuous availability IDC was describing out of the box then NonStop would be the only participant.
There is a lot more to this story, mind you. Transformation to hybrid IT is all the rage right now with HPE among the biggest supporters of hybrids – whether hybrids of servers and private clouds or even hybrids of clouds from different suppliers. However, the word hybrid is pretty scary when you get right down to it and there is no doubt that vendors like HPE recognize the need for them to shine the spotlight on simplifying the transformation. Stated as straightforward as I can – hybrid IT is at best really hard to do and at worst, a path to disaster. For most enterprises the expectation is that vendors will be able to take what they have today, combine with products they would like to have in the future and after a fashion, mold them to where they appear as just a single system albeit a cluster of sorts. Won’t happen – if the Parallel Sysplex aspect to the mainframe has taught us anything at all, adding availability to a system inherently not architected to run 24 x 7 is as much about smoke and mirrors as it is about fulfilling the wishes of individual system programmers who want an opportunity to give it a try.
Hybrid IT is going to amplify the importance of availability. And what it will not do is be a substitute for 24 x 7 and by this I mean, give data center managers a way by which they can opt out of having any part of providing services on the basis of 24 x 7. “We are working with new technology here, mind you, and it is a little flakey so yes, expect outages,” will be the likely response from a data center manager. “We have negotiated with our cloud vendor for backup with redundant compute and storage capacity but we are still working out the bugs and vendors aren’t willing to sign SLAs that demand zero downtime. Can’t be done!” If you read blog posts, as I do, and work with other writers and analysts from across the industry, you will be familiar with how they almost always dumb-down the issue of availability and it really all comes back to no, they can’t tell you how 24 x 7 could be done. As permanent availability, if you like. And the upshot is that they are all discounting the value proposition of 24 x 7 informing all they come in contact with that like real time, there is considerable wiggle room so don’t be too hard on yourselves if you fail from time to time.
As I look back on ten years of blogging and on all that has been addressed in these posts, in a way it seems so strange that I have to write a post about availability and of how no other vendor can provide 24 x 7 support for any of their systems. As surprising as this may seem to many of us, after these ten years we as a community really haven’t done a very good job of extolling the virtues of this really important contribution to the value proposition of a solution. And yet, I am one of those who simply do not like to interact with an application or more likely these days, an app, and find that it is not available for some reason. Hopefully I will still be posting for another ten years and I just have to say it – will I be writing yet another follow-on post to this one and will we still find an industry as blasé about availability as some analysts are today?
I have to admit that this post is following quickly on
the heels of another post just published, a matter of a few days ago. But
tradition remains an important factor; I have written a post to this NonStop
community blog, Real Time View, every August 20 for ten full years. Almost
every August 20, as checking the
archives, there were a few posts that missed the exact date. For me however,
precision is important. It’s rather binary when you think about it. It either
happened or it didn’t. I was reminded about this during a conference call with
a well-known industry analysis organization when the discussion turned to 24 x
7 and the way it was treated; all systems today run 24 x 7, don’t they?
Right now it feels as though I have been working 24 x 7 as I finally return to my office after almost ten weeks on the road. Living and working from our company command center through thick and thin, as we experienced outages of unimaginable proportions, it is good to be sitting behind a desk in a room where there is a sense of permanency. This office isn’t going anywhere nor is my desk and my computer. However, to suggest to you that I have been working around the clock, 24 x7, just to complete this post, may hold some truth given this topic of 24 x 7 has been on my mind for a long time, but in truth, there has been a lot of down time. The office may be a permanent feature of the home but I will still come and go as business needs dictate and that will rarely ever be 24 x 7.
The almost blasé dismissal of the value proposition of 24 x 7 was a wake-up moment for me. An epiphany of sorts, actually! Have we arrived at a point in time where business continuity isn’t all that important anymore? Across the NonStop community the attention we give to the value that comes with being able to run our mission critical applications continuously, 24 x 7, ensuring our customers have access to the applications they need to access any time they want is sacrosanct. After all, with all that we know today about NonStop would it still interest us if it didn’t run 24 x 7? In my last post, New home, new office; time to take stock … I extoll the virtues of scalability and as much as I really like the value scalability brings to any discussion on the value proposition of a system, it still plays a role to aid availability for most stakeholders in the NonStop community as you do not need to stop your operation to add capacity!
In being blasé about 24 x 7 this week, the analyst with whom I was working did think about the value proposition 24 x 7 provides, but then added that truthfully, it is not fated as strongly as it used to be, as even now, just as with real time, the actual time implications for both 24 x 7 and real time are similar. A few minutes here, perhaps an hour once in a while, surely doesn’t matter in the grand scheme of things. No down time means just that – no down time. Likewise, continuously available sounds a lot like permanent availability and readers of posts to this blog will have come to see that I have made reference to permanent availability many times in the past. Marketing types don’t like this term as it has a sense of no wiggle-room about it and seriously, how can you be sure nothing will fail? However, that’s the whole point – NonStop users keep purchasing NonStop systems because they provide permanency to where their clients will always be able to interact with them.
Like my office and the sense of permanency it affords me NonStop systems can be relied upon to be there even if my own access to the applications is casual at best. However, what I am pushing back on these days is terms like “near real time,” as well as “highly available is good enough!” The origins of real time were in missile detection and interception and there could be nothing vague about that. If you didn’t respond in time and neutralize the threat, you simply ran out of time. Nothing real time about pressing the launch button too late to protect a determined target! Again, this is all well-known to the NonStop community and yet, if industry analysts are becoming blasé about this key attribute of NonStop systems and the solutions that have been designed to run on NonStop, then others will follow and it is most definitely the beginning of a slippery slope leading to enterprises associating any value whatsoever to continuous availability.
A couple of years back, IDC did a pretty good job defining different availability levels with availability level four (AL4) representing the very pinnacle of availability. The only vendors to have systems awarded AL4 were HPE with NonStop and IBM with special Parallel Sysplex configurations of the mainframe. The inclusion of IBM really was for appearances sake – IDC just didn’t want a level of availability where only one vendor had achieved that goal. On the other hand, simply trying to build a Parallel Sysplex configuration for a reasonable cost that supported AL4 has proved allusive to almost every mainframe installation. Adding more components – more moving parts, of you like – to an architecture that never intended to support operations 24 x 7 is hazardous at best, but more often proves foolish at worst. If AL4 could only be awarded to vendors with systems that provided the level of continuous availability IDC was describing out of the box then NonStop would be the only participant.
There is a lot more to this story, mind you. Transformation to hybrid IT is all the rage right now with HPE among the biggest supporters of hybrids – whether hybrids of servers and private clouds or even hybrids of clouds from different suppliers. However, the word hybrid is pretty scary when you get right down to it and there is no doubt that vendors like HPE recognize the need for them to shine the spotlight on simplifying the transformation. Stated as straightforward as I can – hybrid IT is at best really hard to do and at worst, a path to disaster. For most enterprises the expectation is that vendors will be able to take what they have today, combine with products they would like to have in the future and after a fashion, mold them to where they appear as just a single system albeit a cluster of sorts. Won’t happen – if the Parallel Sysplex aspect to the mainframe has taught us anything at all, adding availability to a system inherently not architected to run 24 x 7 is as much about smoke and mirrors as it is about fulfilling the wishes of individual system programmers who want an opportunity to give it a try.
Hybrid IT is going to amplify the importance of availability. And what it will not do is be a substitute for 24 x 7 and by this I mean, give data center managers a way by which they can opt out of having any part of providing services on the basis of 24 x 7. “We are working with new technology here, mind you, and it is a little flakey so yes, expect outages,” will be the likely response from a data center manager. “We have negotiated with our cloud vendor for backup with redundant compute and storage capacity but we are still working out the bugs and vendors aren’t willing to sign SLAs that demand zero downtime. Can’t be done!” If you read blog posts, as I do, and work with other writers and analysts from across the industry, you will be familiar with how they almost always dumb-down the issue of availability and it really all comes back to no, they can’t tell you how 24 x 7 could be done. As permanent availability, if you like. And the upshot is that they are all discounting the value proposition of 24 x 7 informing all they come in contact with that like real time, there is considerable wiggle room so don’t be too hard on yourselves if you fail from time to time.
As I look back on ten years of blogging and on all that has been addressed in these posts, in a way it seems so strange that I have to write a post about availability and of how no other vendor can provide 24 x 7 support for any of their systems. As surprising as this may seem to many of us, after these ten years we as a community really haven’t done a very good job of extolling the virtues of this really important contribution to the value proposition of a solution. And yet, I am one of those who simply do not like to interact with an application or more likely these days, an app, and find that it is not available for some reason. Hopefully I will still be posting for another ten years and I just have to say it – will I be writing yet another follow-on post to this one and will we still find an industry as blasé about availability as some analysts are today?
Comments