It was a great pleasure to be part of Money Moves with Deirdre Bolton. The show is about investors looking for alternatives – and hosts a variety of finance and business people. From my perspective, I was trying to convey the message that IT Infrastructure is one of the largest markets on the planet, going through incredible disruption – where the incumbents can do little to prevent it, and new players can emerge quickly and without a huge amount of capital (versus the energy industry which is also going through disruption – but is slow moving and requires a lot of capital).
Below is a short deck on ‘Practical Cloud Economics’, focussed on outlining some of the principles of Cloud Economics, as well as going through common myths surrounding the kind of monetary benefits the Cloud brings.
There are links to some resources within the deck, but here they are again for simplicity, if you would like to read more about the topic:
I get asked about Amazon Web Services pricing literally every day. I suppose it’s a factor of the job I’m in! The AWS Simple Calculator should come with a health warning for anyone who is not already an AWS expert, and the EC2 Pricing page is a resource I personally go to a lot but it’s not a quick reference. We do a good job in CloudVertical of helping Cloud users understand their costs but in terms of just putting together basic, high-level indicative costs for a solution – there isn’t really a simple ‘at a glance’ solution. We do TCO (total cost of ownership) models for customers all the time, and we may launch a public, automated ‘shopping cart style’ TCO and Cloud On-Ramp Calculator in the near future (hint!), but in the mean time, here’s a cheat sheet (PDF) I put together to help with basic, quick, cursory AWS pricing.
A few of the questions it aims to answer at-a-glance:
How much is storage?
How much does an instance cost for a month?
What instance has about 7GB of RAM?
What’s the difference in cost between regions?
How much is data transfer?
How many hours are in a month? How many in a year?
What percentage saving could I get if I bought reserved instances?
Everything is in $USD and based on the US East (Virginia) region right now. If there’s demand for it – I’ll do the same thing in EURO and YEN – based those regions and using currency conversion at the time of creation so it gives a sense of actual cash cost relevant to users in those regions. In the next version, I’d like to add:
Sample Deployments – 1 Web Server, 1 DB Server + expected norms for associated costs; 2 Web Servers, 4 backend Servers; Multi-AZ/Region SMALL and LARGE deployment
Reserved Instances – % of Annual Cost in Upfront Investment, % Hourly Discount (amortized and not), Time Period for Return on Investment.
Spot Instances – Average Spot Instance price per instance and % Saving.
Let me know if there’s any other regular and annoying questions you’d like the AWS Cheat Sheet (PDF) to answer.
Every company has their own ‘stack’ – I thought I would publish the one I use as my reference guide. When talking to enterprises I often say everything that exists in a traditional IT setup, exists in the Cloud – but with different management concerns. For example – it’s not about securing the network so that no data leaves – the data is already ‘out there’ – so in the Cloud it’s about securing the data itself with effective access rights. This creates a ton of opportunities for enterprise startups as the incumbent IT management software is fundamentally broken in the Cloud. It also unfortunately creates a raft of problems for IT departments for whom there is myriad new methods to learn and applications to deploy, to ensure effective service delivery.
I’ve been using this ‘stack’ for a while now – to illustrate that when moving to the Cloud – providers like AWS really only take care of 1 or 2 layers – the rest also need to be taken care of. Very often problems in the Cloud occur when it’s assumed that the IaaS or PaaS provider acts like an outsourcing partner would – rather than simply delivering a commodity service.
There are a few stacks already out there, but I have not found one that comes at this from a systems management point of view – infrastructure deployment to retirement and the lifecycle between. I’m aware that drafting a ‘standard view of the world’ is a contentious thing to do in the Cloud, and so all this is, is *my personal reference framework*. If it’s useful to a few other people, all the better. I use this to outline the various areas that need to be minded when deploying in the Cloud – and I have a list of providers under each heading, and some headings have sub-categories. If anyone wants more info on this, or the players active under each heading – let me know. It’s a ‘Cloud Provider Compendium’ of sorts.
We recently took on Citrix as a strategic investor in CloudVertical (GigaOm covered the announcement, so I won’t repeat it). I wanted to write about why this is a big deal, and why in my mind, for an enterprise startup, working with a strategic investor trumps the other forms of incubator/accelerator.
A bit of background: the Enterprise IT market equates to a 1.7 trillion dollar a year spend, with appalling low effective utilization rates, or extremely high *waste*. Around 20% of physical kit is utilized (usage = average compute usage of capacity available. Like your car, with a capacity of say 130mph, and an average running speed of maybe 30mph – which would be 23% utilization), approximately 30% of virtualized infrastructure on average is utilised (which is mind blowing given the sales pitch for this technology vs. non-virtualized physical hardware), and from our 3k+ users, we know that almost universally from small to large users, about 30% of an IaaS Cloud deployment is utilized. Don’t forget the promise of the Cloud is that in theory it is 100% capital efficient.