- The cloud is not the problem–Part 1: Has it been here all along?
- The cloud isn’t the problem–Part 2: When complex technology meets process…
- The cloud is not the problem–Part 3: When silos strike back…
- The cloud is not the problem-Part 4: Industry shakeout and playing with the big kids…
- The cloud isn’t the problem–Part 5: Server huggers and a crisis of identity
- The cloud isn’t the problem–Part 6: The pros and cons of patriotism
I have been meaning to write a post or three on cloud computing, and its benefits, challenges and eventual legacy. I’ve finally had some time to do so. This series will span over a few posts (not sure how many at this stage) and will focus mainly on SharePoint. In short, I think the cloud is a shining example of innovation, combined with human irrationality, poorly thought out process with a dash of organisational dysfunction. In this first post, I will give you a little cloud history lesson, through the eyes of a slightly jaded IT infrastructure person. To that end, I will try and do the following throughout this series:
- Educate readers to some conceptual aspects of cloud computing and why it matters
- Highlight aspects to cloud computing that are current being conveniently overlooked by proponents (and opponents)
- Look at what the real challenges are, not just for organisations utilising it, but for the organisations providing cloud services
- Highlight what the future might look like from a couple of perspectives
- As always, take a relatively dry topic and try and make this entertaining enough that you will want to read it through 🙂
So let’s roll the clock back a decade or so and set the scene…
In the beginning…
In the height of the dotcom boom of 2000, I took a high paying contract position for a miner-turned-ISP. You see, back then it was all the rage for “penny stock” mining companies – who had never actually dug anything of value out of the ground – to embrace “The interweb” by becoming an Internet Service Provider. Despite having no idea whatsoever about what it entailed to be an ISP, instantly they would enjoy at least a fiftyfold increase in stock price and all the adulation of those dotcom investors who actually believed that there was money to be made.
Lured from my stable job by the hubris-funded per-hour rate and a cooler job title, I designed and ran an ISP from late 1999 till late 2004, doing all things security, Linux, Cisco and Microsoft. Back then, the buzzword of choice was “hosting”. Of course, the dotcom bubble popped big time and the market collapsed back to cold hard reality pretty quickly. Like all organisations that rode the wave, we then had to survive the backwash of a pretty severe bear market. Accordingly, my hourly rate went down and our ISP sales guys dutifully sold “hosting solutions” to clients that were neither useful nor appropriate. The best example of this is when someone sold a hosted exchange server to a company of 300 staff with no consideration whatsoever of bandwidth, security and authentication (remember that this was the era of Exchange 2000, immature Active Directory deployments and 1.5/256 megabit ADSL connections).
We actually learnt a lot from dumbass stuff like this (and we went through a seemingly endless number of sales guys as a result). By the end of the journey, we did some good work and had a few success stories. The net result of riding the highs and lows of the dotcom boom, was my conclusion that if you had a public IP address and a communications rack with decent air conditioning, you were pretty much a hosting provider.
Then in 2004 I took a different job with a different company. They hired me because they had just acquired a fairly well-known “hosting provider” who had gone through some tough times. I was tasked with migrating the hosting infrastructure – and the sites hosted on it – to the parent company premises and integrate it with the existing infrastructure. So imagine my shock when on day one, I arrive onsite to see that the infrastructure of this hosting provider was essentially a store room, full of clone PC’s with panels removed, sitting in a couple of communications racks, with a cheap portable fan blowing onto it all to keep it cool and with no redundant power (in fact one power cord was sticky taped to the floor and led out the room to the nearest outlet). As it happened, some very high profile websites ran on this infrastructure.
This period I describe as “my bitter and twisted days” as I had a limited time to somehow migrate this mess to the more robust infrastructure of the parent company. This was the period where I became a bit of an IT control freak and used to take a dim view of web developers who dared to ask me a dumb question. I also subsequently revised my view of hosting. I decided that if you had a public IP address and a comms rack with completely crap air conditioning, you were pretty much a hosting provider. After all, when you access a website, did you ever stop to consider where it physically might reside?
…and henceforth came “the cloud”
Before SharePoint 2010 came out, I used to do talks where I put up the SharePoint 2007 pie and asked people what buzzword was missing. Many hands would rise and the answer was always “cloud”. Cognisant of this, I redrew Microsoft’s marketing diagram to try and capture the essence of this this new force in enterprise IT. I suggested that Microsoft would jump on the cloud big-time with SharePoint 2010. How do you think I did?
As it turned out, Microsoft for some reason opted not to use my suggested logo and instead went with that blue Frisbee with fresh buzzwords to replace the 2007 ones that had reached their saturation point. Nevertheless, the picture above did turn out to be prophetic: The era of the cloud is most definitely upon us, along with the gushing praise that often accompanies any flavour of the year technology.
Now in one sense, nothing much has changed from the days of web hosting. If you have an IP address with a webserver on the end of it, you can pretty much call yourself a cloud provider. This is because at the end of the day, we are still using the core ingredients of TCPIP, DNS, HTTP, communications racks and supposedly good air conditioning. When you access something in “the cloud”, you have no visibility as to the quality of the infrastructure on the other end. For all you know, it could be a store room being kept cool with a dodgy fan and some sticky tape :-).
But while that’s a cynical view, its is also naively simplistic. Like all fads that come and go, things are always changed as a result. The truth is that there has been changes from the days of web hosting that will change the entire face of IT in the coming years.
The major difference between this era and the last is the advancement in technology beyond those core ingredients of TCPIP, DNS and HTTP. Bandwidth has became significantly cheaper, faster and more reliable. Virtualisation of servers (and services) not only gained momentum, but is now a mature technology. My own evidence for this fact is that I haven’t put SharePoint web front end servers onto non-virtualised infrastructure for a couple of years now. Add to that the fact that the tools and systems that we use to build web solutions are now much more powerful and sophisticated. As a result, “cloud” applications now reflect a level of sophistication and features way beyond their web based email origins. Look at Office 365 as a case in point. Microsoft have bet big-time on this type of offering. I’m sure that most architectural diagrams currently drawn all over Microsoft whiteboards for SharePoint vNext, will be all about reworking the plumbing to create feature parity between on-premise SharePoint and it’s cloud based equivalent.
It’s interesting stuff indeed.
Now, perhaps because I had an ISP/hosting ringside seat, I could see all of this happening way back in 2000 – more than a decade ago. Not only could I see it, I experienced the pain of early adopters trying to do it (witness the example of the hosted Exchange 2000 “solution” I started this post with). But a decade later, cloud based infrastructure now realises the sort of capabilities that I was able to foresee in my ISP days. We have access to unlimited storage and scalability. With it, I can save massive time and effort to get complex systems up and running. In this fast-moving age we find ourselves in, being able to mobilise resources and be productive quickly is hugely important. Recognising this, companies like Amazon, Google and Microsoft leverage their incredible economies of scale, as well as the sheer depth of technical expertise to make some rather compelling offerings. Bean counters (i.e. CFO’s and CIO’s with tight budgets) suddenly realised that the cost to “jack-in” to a cloud based solution is way less costly than the traditional manner of up-front costs of hardware, licensing, procurement and configuration.
The cloud offers minimal entry cost because for the most part, it is based on a pay-for-use model. You stop paying for it when you stop using it. Buying servers are forever, but the cloud is apparently not. Furthermore, the economies of scale that the big boys of the cloud space offer, usually far exceeds what can be done via internal IT resources anyway. This extends past sheer hardware scalability and includes security, reliability and performance monitoring. As a cloud provider customer, you will not just expect, but assume that companies like Microsoft, Amazon and Google can use their deep pockets to hire the best of the best engineers, architects and security practitioners. Organisational decision makers look increasingly longingly at the cloud, in the face of internal IT costs being high.
Even the most traditional on-premise IT vendors are getting in on the act. Consider SAP, previously a bastion of the “on-premise” model. Their American division just shelled out US$3.4 billion to buy a cloud provider called SuccessFactors (3.4 billion = 50% premium to SuccessFactors share price.) Why did they do this? According to Paul Hamerman (the bold areas are mine).
“SAP’s cloud strategy has been struggling with time-to-market issues, and its core on-premise HR management software has been at competitive disadvantage with best-of-breed solutions in areas such as employee performance, succession planning and learning management. By acquiring SuccessFactors, SAP puts itself into a much stronger competitive position in human resources applications and reaffirms its commitment to software-as-a-service as a key business model.”
If that wasn’t enough, consider some of Gartner’s predictions for 2012 and beyond. One notable predictions is that by year-end 2016, more than 50 percent of Global 1000 companies will have stored customer-sensitive data in the public cloud. Closer to home for me, I have a client who has a ten-year BHAG (known as a Big, Hairy Audacious Goal). While I can’t tell you what this goal is, I can tell you that they have identified a key success metric that currently takes them around 12 months to achieve. Their BHAG is to reduce this time from 12 months to 4 weeks and achieve this within a decade. Essentially they have a time-to-market issue – similar to what Hamerman outlined with SAP. By utilising cloud technology and being able to procure the necessary scalability at the click of a button and the swipe of a credit card, I was able to save them one month almost straight away and make a massive inroad to their organisation-wide strategic goal.
So it seems that in the rational world of key performance indicators and return on investment, and given the market trends of large, mainstream vendors going “cloud”, it would seem that we are in the midst of a revolution that has an unstoppable momentum. But of course, the world is not rational is it? If it were, then someone would be able to explain to me why the US still uses the imperial system given that every other country (save for Liberia and Myanmar) has now changed to metric (yes my US readers, the UK is actually metric).
The irrational road ahead…
In this first post I have painted a picture of the “new reality” – the realisation of what I first saw in 2000 is now upon us. While this first post might sound like gushing praise of all things cloud, rest assured that this is not the case. I deliberately titled this post “the cloud is not the problem” because we are going to dive into the seedy underbelly of this brave new cloudy world we find ourselves in. My contention is that cloud computing is an adaptive challenge, which by definition, questions certain established ways of doing things. Therefore it has an effect on the roles, beliefs, assumptions and values behind the established order. In the next post or three, we are going to explore some of the less rational sides of “the cloud” at a number of levels. Furthermore, the irrationality often tends to be dressed up as rationality, so we have to look behind the positive and negative straw-man arguments we are currently hearing about, to what is really going on. Along the way I hope to develop your “cloud computing strawman argument” radar, so you can smell manure when its inevitably dished out to you 🙂
The general breakdown of this series will be as follows:
I’ll start by chronicling my experience with Microsoft’s new Software as a Service (Saas) offering: Office 365, as well as Amazon’s Platform as a Service Offering (EC2). Both are terrific offerings, but are let down by things that have nothing to do with the technology. From there we will move into looking at some of the existing roles and paradigms that are impacted by the move to cloud solutions, and the defence mechanisms that will be employed to counter it. I’ll end the series by taking a look at the cloud from a longer term perspective, based on the notion of systems theory (which despite its drop-dead boring sounding premise is actually quite interesting).
Thanks for reading