Why use public cloud
If you haven’t already read our “Which cloud solution is right for me” article then that’s definitely somewhere to head after reading this 😉
Before we get into the main bulk of this article it’s probably worth mentioning that there are two *main* cloud options to chose from: private cloud and public cloud.
Private: Using a private cloud for your hosting needs gives you access to a certain amount of compute power for a set price. This means you’re paying to reserve a set amount of system resource (a slice of a bigger computer normally). While it does allow for a steady bill each month, it also means you’re paying for the maximum amount of resource you think you could use, and then not using it all the time.
Public: Using a public cloud properly ensures you only pay for the resource you use. It also means a sudden surge in visitors to your site/application is handled by the platform. A public cloud has features that can automatically replicate/create new instances of your environment to make sure demand is met in high-traffic times.
If that all made sense then you probably don’t need to be reading this article, but if you’re still curious…
Let’s put this into a (simplified!) scenario
Your widget selling website has seen a sudden increase in people visiting one evening – maybe due to an ad campaign on the telly – and the load is nearing your 80% panic threshold. Thankfully you employed a skilled cloud consultancy to build the site’s infrastructure, so once the 80% threshold is reached, your public cloud platform knows to automatically build n identical servers/services to help out.
This process takes a matter of seconds and before you know it your website has additional resource, quite possibly in the form of two new web nodes. Huzzah. As the amount of concurrent visitors increases, so does the amount of underlying web nodes (to meet demand). When the number of visitors starts to decrease, your public cloud platform will keep an eye out for its minimum load threshold (say, 30% load) and will start to remove web nodes that aren’t required.
This will continue until you’re left with two or more nodes (depending on your requirements). All this happens in the background and doesn’t affect your performance. It also means that throughout this entire example, the widget selling company only ever paid for the resources it was using, rather than paying for the maximum amount the company thought they might need at their busiest time.
It’s probably worth noting at this point that assuming a maximum amount of resource in a private cloud can be tricky: if you over-commit to resources you’re wasting a lot of money on compute power that’s sitting idle. And if you under-commit, then when the time comes and all those visitors arrive on your site to buy widgets, your site will fall over and nobody will be able to buy widgets. Nightmare!
As mentioned previously, this is a simplified version of what can be achieved 🙂
Back to business
Due to the immense amount of assets that the big three (Amazon, Microsoft and Google) have in their cloud offerings, they are able to offer a surprisingly large amount of resource for a relatively low price. Considerably low when compared to purchasing/renting physical servers from hosting providers. These economies of scale are on a different level to anything else offered in the hosting market. Not only that, but now that Amazon has competition in this industry there has been a very prominent price war emerging. This means that there’s a clear “race to the bottom” when it comes to the prices they offer their customers (this is still happening frequently now).
The large amount of assets the big three have means they are able to spread services all around the world. So not only do you get the economies of scale, you get easy and affordable fault tolerance (should that be something you require). You can also use this to your advantage if you had a client-base in another country to the one you operate it. Public cloud solutions are geo-diverse and as such having presence in two, three or many countries becomes a lot simpler.
Although finance and efficiency are definitely the ma in reasons to switch to a public cloud environment, there are other strong factors…
Public Clouds offer a number of “on-demand” services such as Database, CDN, etc – that take the management & scaling of those services away from the end user. A lot of these services would originally require being on a fully functional server to operate correctly, but with the public cloud, you would only need the service itself, once again, freeing up money from unnecessary resource usage. These on demand platforms can get quite granular (eg, SQS, RDS, ElastiCache) so I won’t go into too much detail here, but imagine having all these additional features at your disposal that improve efficiency, flexibility and performance; pretty impressive.
The on demand services, along with the more standard features of a public cloud offering mean you have a very granular level of control. Every aspect of your hosting environment can be controlled, limited and measured exactly how you’d like. Using a specialist cloud consultancy for something like this would be favourable for organisations that don’t have the expertise in house as there’s a lot to learn and the potential to be continually enhancing your platform.
On the topic of granularity, there’s also the levels of access control across your cloud services. If you were to use a company like Steamhaus to manage your public cloud environment, you could still be the overall owner of your services, but grant access to us so we can manage things without the worry that we could lock you out.
There are a lot of assumptions and misconceptions about the public cloud which tend to mean organisations don’t want to take the plunge into or risk trying a new environment. There’s also the time factor where if a private cloud environment is working most of the time for them, why waste resource understanding something when there are more pressing matters?
Sure, that makes sense, but how about if you could use a consultancy that has a wealth of knowledge to do that hard bit for you? Especially if it meant your platform would become more cost-effective and/or more resilient.
As a side note…
It’s worth noting that some sites or applications don’t need – or benefit from – the full feature-set and instant scalability of large-scale public clouds. Some sites and applications aren’t designed for the architecture of a large-scale cloud provider, instead needing to have high availability handled by the underlying hosting platform. There is also the factor that some would just prefer their data to be hosted locally, managed by a company they can visit. That’s why we created SteamCloud.
There is great comfort in knowing how much your hosting is going to cost every month, which is exactly what you get with SteamCloud, as well as high-availability whilst being completely secure. It has the best hardware and a well thought-out network topology, based on years of experience with virtualisation in a hosting company environment. It’s based on VMWare, and located right here in sunny Manchester, in our local Partner Datacentred’s Tier 3 facility in Manchester’s MediaCity.