Forget about Saas (software-as-a-service). The newest next big thing is PaaS (platform-as-a-service).
As the name implies, the concept behind PaaS is that an entire platform-specific technology stack as a service. As with SaaS offerings, buzzwords like 'cloud' and 'scalability' are typically thrown around when pitching the benefits of PaaSes.
Thanks to Salesforce.com's $200m+ acquisition of Ruby on Rails PaaS provider Heroku late last year, everyone is lining up to offer [insert platform name] PaaS solutions. And investors are flocking.
Case in point: PaaS startup PHP Fog, which earlier this year raised close to $2m, launched publicly last week. As its name hints, the company is basically a Heroku for PHP.
Its value proposition: it's a "rock-solid cloud platform for PHP." Yet there's a lot PHP Fog lacks. You can't install PHP extensions of your own, and in many cases, popular, commonly-used extensions aren't even available by default. To boot, PHP Fog has been hacked, and it's built on the same Amazon EC2 infrastructure that took down Heroku recently.
What it does offer: a one-size-fits-all solution that offers some niceties (such as a Varnish-based caching layer). And hefty pricing. A Platinum account, which costs $249/month, offers a MySQL database with just 20MB (yes, megabytes) of storage space. Yet somewhat amusingly, PHP Fog says that all its accounts come with a "scalable MySQL database."
To be fair, PHP Fog is hardly the least presentable PaaS offering out there; I'm simply discussing it because I've seen the LAMP stack up close in some very demanding environments and it is a fairly recent entrant into the space.
The reality is that the majority of these PaaS offerings arguably don't do much for low-end and mid-range customers (developers, small and medium-sized businesses, etc.) who will never need to "scale" in any meaningful way.
For these types of customers, most PaaS solutions are also ridiculously overpriced when compared to dedicated offerings that would provide ample horsepower (and then some).
For larger customers with high-volume and mission critical applications, these PaaS offerings offer nowhere near the level of control and configurability needed. Note that this isn't just true in the software stack, it's true in the hardware stack. Case in point: Amazon EC2's disk I/O performance is simply not going to cut it for certain applications at scale.
At the end of the day, any company that is serious about application performance, scalability, security and extensibility needs to have full control over its technology stack and infrastructure. Period.
While this doesn't mean that there isn't a market for PaaS offerings, the truth of the matter is that the way they're being sold (read: production-ready, scalable, cost-effective solutions for applications of all sizes) is simply not accurate.
In many cases, the primary reason companies are buying into these solutions is that they've given their developers far too much control over the technology stack, a mistake that many companies will eventually recognize the hard way.
Yes, it's nice to believe that you can build an application and scale it instantly and cheaply with a few clicks, but unicorns are nice too. It doesn't make either real.