Follow the Australian Telecommunications scene NEWSLETTER- FREE TRIAL
 | |
|
Technology news and Jobs 
The Linux distillery Google closed source app engine does evil
The Linux distillery Google closed source app engine does evil | Google closed source app engine does evil |
|
| by David M Williams | |
| Wednesday, 16 April 2008 | |
|
Page 3 of 3 The biggest problem with Google’s App Engine is vendor lock in. You’re stuck. You want to later move platforms? Maybe your database code will be pretty portable but the transparent scalability won’t be. You want to export your data store? You can’t; there’s no export or dump features.Featured Whitepaper
5 Best Practices for Smartphone Support
Lock in is not trite; the touted custom APIs – which include user authentication and management – must be replicated elsewhere and these are not simple undertakings. The App Engine is not open source – and perhaps the reasons for this can be understood; if Google divulged the deep secrets of their sandbox wiley characters may seek to circumvent it. Yet, by the same token, pragmatic honest developers have no certainty about the security of their apps because they have to take the platform’s security on face value. By all means, the Google App Engine is a marvellous achievement and undertaking and a generous gift to the web – but any developer or startup company seeking to build a presence on it should recognise from the beginning they are going to be committing to the platform. Future migration away will not be a trivial undertaking. And while about 5 million pageviews per month are promised, there’s no service level guarantees. Maybe there ought not be; it is at no cost after all. However, do you wish to leave your code in the hands of no uptime assurance? I’ve seen just today how the SDK and web assistant themselves have failed me. Commercial organisations need especially to contemplate this. Google are holding their cards close to their chest about the metrics they think they can achieve. It won’t be until a hefty sample size of apps are installed and running that they can reasonably state how their engine holds up. Similarly, it won’t be known for the near future what the pricing plan will be for more than 5 million pageviews. The risk is you will start with the free service, and find out too late that it won’t meet your load and a better plan is either too expensive or, as at present, non-existent. Your code will have to be rewritten if you wish to leave and your data won’t come with you without serious coaxing. If you’re a developer you must put considerable thought into whether you wish to host with Google, whether this offering is right for you. At this time the roadmap is still too unclear to make such determinations with clarity. |
Tags
| < Next story in category | Previous story in the category > |
|---|
