Webreakstuff's blog on design, development and strategy. Click here to subscribe.

Service dependency on the new web

Fred Oliveira on October 27, 2005 Comments (14)

Web Services Reading this post on the O’Reilly Mac Developer Center got me thinking about something we’re slowly falling into: web-service dependencies. Most new web applications are, in one way or the other, dependant on third parties, and the funny thing is noone seems to worry much.

Looking at the plethora of apps that use Google maps (and this is an example), is cause for concern. If google decided to drop support for Google maps, what would all the developers do? What would happen if Flickr decided to stop allowing people to use their photo database? What if?

Fact of the matter is, after service outsourcing and personal outsourcing, we’re seeing a new age of web-service outsourcing. One with no regulations – only expectations and hopes. Everything is based on trust, and trust sometimes fails. And the problem here is that even with web-services as a liability, there’s no fallback mechanisms, no alternative routes, no “competitor service” that can be plugged into an app in the timely manner live web 2.0 applications require.

This proves that purely mash-up based applications have small foundations, and like a house, with no foundations they may fail to resist should the unexpected happen. What will we do when services turn (if they do turn) greedy? How will we change if outsourced services change or close, since we’re seemingly so dependant?

A question for developers: Does your app depend on a third party service? Do you have an escape route or alternative? Should someone go crazy on a board meeting and decided to close a webservice you depend on, would you close doors? I would definitely love to hear some opinions on this subject.

View/add comments on this entry

Comments on this post

Hooman Radfar

Fred, I could not agree with you more whole-heartedly. Remix development is at a nacscent stage. However, as more developers leverage existing services, this issue will become far more prominent. Currently, existing web service providers maintain sites in a manner consistent with developer expectations in order to maintain reputation. However, as new services our built by under capitalized entities, how can we have that same level of trust? If web service adoption and corresponding development efforts continue to grow, I believe that this issue will come to light. Great post.

Pablo

I´m agree with you Frederico that a complete business model based on other people´s service is very risky. Excellent post.

Bob Aman

I am writings services that rely on other services, and this has me very concerned, yes. However, I simply cannot afford to implement all functionality myself without relying on any kind of externality, especially if the information in question is subject to frequent change.

So I try to keep a backup plan handy, and I always make sure that any call out to a web service API has some sort of error handling, preferably with fall-back data, either in the form of caching or intelligent guesses (as in the case of geocoding, for example) or some combination of the two. So if network issues or “crazy decisions in board meetings” cause an interuption to the service, I’m not dumping errors. And in fact, I’d hope most users wouldn’t even notice anything amiss.

And then I hope that buys me enough time to find a replacement source of information if required.

Peter Foley

Great post. I guess conventionally you would expect the emergence of for fee service level agreements, which is common in business to business relationships, but where the relationsips are based more on individuals and/or smaller groups this model may not be as practical.

It would be very interesting if alternatives appear on the market, and applications are developed for failover, ie google maps is unavailble, use XYZ maps as an alternative. With the development of standards this may be possible.

Scott Motte

Very interesting post. I don’t ever see these services disappearing though. In fact, I see the whole situation as reminiscent of what Microsoft did.

Microsoft’s OS became the platform for the PC stage, and everyone had to buy Windows in order to “make things.”

Now Google, and Yahoo’s services are becoming the platform for the Web stage, and everyone has to see their advertising in order to “make things.”

Pete Cashmore

“If google decided to drop support for Google maps, what would all the developers do?”

More to the point: if Google starts putting ads in Google Maps, what will the remixers do? And will Google share revenues with other sites? Judging by http://www.trulia.com, a real estate site which is talking to Google about revenue-sharing, it looks like they might.

Of course, Google Maps does have a “competitor service” – MSN Virtual Earth. And – perhaps in acknowledgement of Google’s dominance of map mashups – the Microsoft guys have hired some map hackers. These hackers have created some code which lets you instantly switch your Google Map for an MSN Virtual Earth map, so the remixers will now have a choice. Likewise, if Flickr suddenly restricted access to their database, I expect its competitors would seize upon the opportunity and create an easy way to switch.

James Kahler

Good work. Risky? Yes. But so are these start-ups by nature. Basically, I believe what we are seeing is a bootstrapping of not only companies, but of a business eco-system. Of course many of businesses will fail, which will lead to other failures. Others may wall off their gardens (services), putting still others out of business. At the end of the day we will see a small set of winners who have aligned with others who are winners, but at this stage of the game, it may be impossible to differentiate between winners and losers on which to base a technology.

But without this leap of faith, I don’t think this ecosystem gets off the ground.

WeBreakStuff » Web applications: APIs, use but be ready

[...] Service dependency on the new web, Published October 27 Public APIs – Arghh, Published January 11th on Forest and the Trees Web 2.0 and the Drive-by upgrade, Published October 24 on O’Reilly Weblogs [...]

benningblog » Grizzly Man

[...] Service dependency on the new web, Published October 27 Public APIs – Arghh, Published January 11th on Forest and the Trees Web 2.0 and the Drive-by upgrade, Published October 24 on O’Reilly Weblogs [...]

services

I felt good about this post. It confirmed for me some of the things I’ve been thinking about.

Webified Desktop Apps vs Browser-based Apps « Bhprojekt’s Weblog

[...] Fred Oliveira of WeBreakStuff nicely put it: “…after service outsourcing and personal outsourcing, we’re seeing a new age of [...]

KraKerDesign

KraKerDesign

http://www.krakerdesign.com

web tasarım web hosting reseller domain hosting windows linux

webabi

WebAbisi.Com

links for 2006-09-09 - Buzzsonic.com

[...] Webreakstuff » Service dependency on the new web [...]

Something to say?