“As a Service” services (service2?) generally allow younger companies to scale quickly and efficiently. A lot of the hassle is abstracted away from the pain of implementation, and they allow start-ups to focus on the key drivers of any company – product quality and product availability. For less than the cost of proper infrastructure investment, you can have highly-available, fully distributed, buzzword enabled things at your fingertips to start running wild with.
However, “as a Service” providers feel like they’re filling a short-term void rather than building long-term viable option for companies. Here’s why.
1. Cost
The main driver of SaaS is that there’s lower upfront costs. But it’s a bit like the debit card versus credit card debate; if you have the money you can pay for it upfront and never worry about it again. If you don’t have the money but need it now, then credit is the answer – and the associated continued costs.
For start-ups, a perceived low-cost model is ideal at first glance. With that, there’s the downside that you’ll be paying out of your aaS for the rest of your service with them, and moving out of the ecosystem that you thought looked so robust 4 years ago will give the sys admin that you have to hire in to fix it nightmares.
Cost is a difficult thing to balance, but there’s still companies still happily running on SQL Server 2005 without any problems; a high upfront cost normally means that it’s going to stick around for ages (you’ll make it work!). To be honest, for most small businesses, investment in a developer who can stitch together open source technologies to suit your needs will be better than running to the closest spangly Service provider.
However, aaS does mean you don’t need System Administrators stressing about ORM-generated queries.
2. Ownership of data
An under-discussed but vital issue that lies behind the aaS movement is the ownership of data, and what this means to companies. How secure are the bank details of your clients? How does the aaS provider secure against attacks? Where does this fit in terms of compliance?
To me, the risks associated with giving your data for another company to keep is too high to justify, even if it’s backed up by license agreements and all types of unhackable SSL things (#Heartbleed). After all, a bank is more appealing to thieves than a safe behind a picture in your living room. Probably*.
As a company, regardless of size, your integrity is all. I think you should own that.
3. The Internet as kingmaker
We once had an issue at the Packt office where, during a desk move, someone plugged an Internet cable (that’s the correct term for them, right?) from one port to another, rather than into their computer. The Internet went down for half the day without anyone really knowing what was going on. Luckily, we still had local access to stuff – chapters, databases, schedules, and so on.
If we were fully bought into the cloud we would have lost a collective 240 man hours from one office because of an honest mistake. Using the Internet as your only connection point to the data you work with can, and will, have consequences for businesses who work with time-critical pieces of data.
This leaves an interesting space open that, as far as I’m aware, very few “as a Service” providers have explored; hybrid cloud. If the issue, basically, is the Internet and what cloud storage means to you operationally and in terms of data compliance, then a world where you can keep sensitive and “critical” data local while keeping bulk data with your cloud provider, then you can leverage the benefits of both worlds. The advantages of speed and lack of overheads would still be there, as well as the added security of knowing that you’re still “owning” your data and your brand reputation.
Hybrid clouds generally seem to be an emergent solution in the market at large. There are even solutions now on Kickstarter that provide you with a “cloud” where you own your data. Lovely. Hell, you can even make your own PaaS with Chef and Docker.
I could go on. The quite clear popularity of “as a Service” products means there’s value in the services they’re offering. At the moment though, there’s enough problems inherent in adoption to believe that they’re a stop-gap to something more finite.
The future, I think, lies away from the black and white of aaS and on-premises software. There’s advantages in both, and as we continue to develop services and solutions that blend the two, I think we’re going to end up at a more permanent solution to the argument.
*I don’t actually advocate the safe behind a picture method. More of a loose floorboard man myself.
From 4th-10th April, save 50% on 20 of out top cloud titles. From AWS to Azure and OpenStack – and even Docker for good measure – learn how to build the services of tomorrow. If one isn’t enough, grab 5 for just $50! Find them here.
