fdisk.it

The ever growing world of mobile apps, is Cloud the answer?

The boom in mobile apps has often been seen as a liberalising one, a market open to anyone, or any group, with the ability to put together a ‘killer app’. It’s in some ways reminiscent of the early days of videogames, with classic titles such as Manic Miner or Dizzy, designed and programmed by one or two young upstarts rather than today’s teams of dozens. App stores have meant a reawakening of these “bedroom programmers”, giving them a place alongside well-financed start-ups and long-established software developers.

The reality isn’t quite so rosy however, App stores can be cruel places for the unknown app: if you want your app to be known, you may need to go through an opaque approvals process and device manufacturers can limit how you can make money; or, on the other hand, the lack of approvals on others means your app can get lost in a sea of poor quality alternatives.

So what can an app developer do to avoid these traps? One way to circumvent the app store is to build your app in the cloud – Software as a Service. By using a mobile device as a way to access the app through the browser, developers have more freedom to create what they want, and monetise it how they want. Updates become easier and more frequent as they happen in the cloud and don’t have to be installed on the device. By using HTML 5 standards, developers can also avoid doubling up, or worse, on work – one app will work on iOS, all the varieties of Android, Windows Mobile, and even less common mobile operating systems.

There are also the advantages of the cloud itself. Limited by their size and battery life, mobile devices lack the processing power of desktop PC or laptops. Anything too complex can bring the device to a crawl, and the wide variety of devices available means developers looking for the widest possible audience must develop for lower-spec devices rather than taking advantage of the power of the latest handset. Applications that need a large database, or need to make a large amount of calculations, can use the processing capabilities of cloud computing to create apps that perform complex tasks without overburdening handheld devices.

Using the cloud may seem the obvious choice for app developers, giving them the freedom to do what they want, charge how they want, and make things as complicated as required. But you the developer must make sure all the bases are covered. Cloud hosting means additional cost and the supplier of these services must be able to provide SLAs that cover any risk of downtime, and your customers connection must be reliable in order for them to use your app. And while app stores aren’t necessarily as open or as democratic as we’d like, they do provide some exposure for new and well-reviewed apps.

The decision to make your app Cloud-based or not is an important one that needs to be made at the early stages of app development, and will depend on what you are trying to achieve. Big organisations can afford to flip-flop between HTML5 and native apps – Google has changed its mind more than once – but smaller organisations need more concrete plans to avoid the expense and time wasted switching from one to the other.

The ‘Slashdot effect’

It’s an experience common online – someone sends you a link via email or Twitter or Facebook, perhaps to the newest social networking tool, to an interesting news story, or a photo of a cute kitten. You click on the link and… nothing. Whatever the link is, it’s been too popular for its own good and can’t handle the heavy traffic.

This was known as the ‘Slashdot effect’, after the technology website where links would appear then almost immediately go down because of the attention. Slashdot’s popularity may have waned, but sites such as Reddit and Digg, or popular tweeters such as Stephen Fry or Ashton Kutcher, can make a website unreachable. What’s actually happening here? The number of visits to a website isn’t really the problem. The problem is concurrency – the number of visitors per second. A good way to imagine the problem is a turnstile at a football stadium. The stadium may be able to hold tens of thousands of people, but only a very limited number can fit through the turnstile at once. Try to fit too many through the turnstile, and the whole system breaks down. When hackers bring down websites in a ‘denial of service’ attack, this amounts to the same thing.

Cloud is often seen as the ‘silver bullet’ to cope with high traffic loads, but sites need to be fully optimised from the ground up with that in mind – adding cloud as an afterthought to cope with traffic often isn’t possible. Businesses need to make sure that their website is optimised, using the right databases and correct programming, to take advantage of Cloud hosting.

More importantly, businesses need to consider what they actually need. Sometimes high loads come unexpectedly, but often they are as a result of planned activity – product launches, events or celebrity endorsements. In these cases, your hosting provider can plan for this extra load – adding capacity as it is needed and scaling it down when it isn’t.

Finally, there is always the option of simply not meeting the load, if it’s unnecessary. For example, a ticket agency, or international corporation may need to be online every second of every day, but a few minutes of downtime may not affect your business too much, and the extra cost may be unjustifiable. In this case, it’s important to fail gracefully – a simple page assuring your audience that you’ll be back soon is better than any error message.

Cloud is the future, deal with it!

Contrary to popular opinion this is not aimed at the English ‘Summer’ we’ve had so far this year, moving on…

Over the last twelve months I’ve been having daily conversations on the topic of Cloud Hosting with colleagues, customers, journalists, friends and leading players in the market on the subject along with speaking at various events.

The conversation goes something like this (pardon the stereotypes!):
Continue reading →