I had the pleasure of attending the TECH.unplugged event in London this week, hosted by Enrico Signoretti. This was an independent event with a number of technology analysts from around the world providing many interesting opinions on a wide variety of subjects.
One session in particular which fired up my thinking about the future of the tech industry was from Stephen Foskett. For those who don’t know, Stephen organises and hosts the Tech Field Day events in the US (techfieldday.com). As an aside, these events are worth keeping an eye out for to see deep dive discussions with vendors over their latest technological offerings. This talk at TECH.unplugged was one which really twigged with me, and over the last few days I have mustered the thoughts below together.
Stephen’s talk was around the idea that Cloud Computing, everyone’s favourite buzzword, will become just a part of computing. This was illustrated by numerous points in the development of computing through the ages, which never went away, despite them falling out of favour and becoming unsexy in the tech world. These include:
The mainframe – still in use today, and still very much alive. See the new IBM z13 mainframe, this was only released in Jan 2015, but shows that there is still development, and iteration in mainframe products today. These are still used across many industries, and although the IT industry is not focussed on these, they are still a crucial component of what we deliver
Tape backup – people would love this to go away I am sure, tape has been around for decades, and presents a cornerstone in backup and data retention strategies, from small to large organisations. In a time where we can fit many terabytes of information on a chip the size of a fingernail, why are we still using tape? The answer is that tape continues to present the two same advantages it always has: it is cheap, and it is reliable. As much as we see $/GB prices fall and fall with magnetic disk, and now with SSD, tape still kills it in price
Physical servers – sure, virtualisation is king in the data centre today, and for the last 5-10 years, that stuff left running on bare metal tin, without a hypervisor, better have a good reason for that or it is going to get P2Vd at some point.
So how does Cloud fit in here? The industry heralds crow their message over and over again: “The data centre is dead, long live the Cloud”, and those of us still supporting physical tin are hoping and praying that the prophecy does not come true. We hear tales of the developers rallying under the DevOps banner, with Docker, Vagrant and Puppet as their drawn weapons, ready to rise up against the slow traditional infrastructure and move our business’ most critical workloads to AWS to allow them the freedom of rapid, agile development they need.
But realistically, will this spell the end to all we have built in our data centres, for all of AWS’ £1bn quarterly earnings, and their 100% year on year growth reported this week, will this actually kill the data centre. Well no doubt, there are applications, many of them being written today, which are what VMware calls ‘Cloud Native Apps (CNA)’, that is applications designed to run at scale, to be designed around the micro-services model, and to deliver scale up and down on demand. Applications utilising continuous integration and delivery models, to allow hundreds of code releases each day, delivering what the business needs in near-real time.
Well it seems that this is a new part of the puzzle, filling a hole which we in the data centre business did not realise was there, but one which developers leapt to utilise. We have tried to deliver, but alas we were too late. We tried to deliver solutions like vCloud Director, like OpenStack, but it seems these products were not quick enough, or too clunky, or just too damned difficult to install. So our businesses are already using cloud services, and they are seeing just how great it is. We are losing control of our IT.
But all is not lost, as far as our workloads go, there are no doubt applications perfectly suited to the cloud (for some businesses, at least). We use ServiceNow at my current workplace as our CRM software, and since moving to the cloud the application is faster, more available, and updated far more often.
There are no doubt applications that some businesses do not want to put in the cloud though, and despite arguments from both sides, the feeling that data is safer in your on-premises data center, than it is in Amazon’s or Microsoft’s, is still commonplace. Cloud advocates insist that one day all applications will be in the Cloud, I don’t think this will be the case. As Stephen illustrated in his presentation, Cloud Computing will be but one tool that the IT industry has to deliver the services end users need.
VMware’s release of Photon is a good analogy for the way I feel that software delivery will go, with private cloud getting software and hardware offerings which mirror what is going on in public cloud. This will give those IT organisations who don’t want, or are not allowed, to put their applications and data in the cloud, the same tools in-house that developers are crying out for. So we should embrace tools like this, encouraging vendors to release more tools like Photon/Lightwave, which gives us a secure, on-premises approximation of what our developers cry out for, and most importantly, learn to implement, use and support these tools, so that we can be part of the Cloud revolution, and be ready to bring these tools into our toolset.
The overarching message from Stephen’s presentation though, was that Cloud is coming, and soon it will just be another part of what we do. As technology advocates, we learn new tech all the time, and the DevOps movement, Cloud, containerisation, whatever it may be, it is all just more new tech to learn, so don’t be afraid; learn it, do it, master it. We are IT people, that’s what we do.