Sign in or 

Note: I define "Software Architecture" as the architecture of systems that involve software systems; this includes Enterprise Architecture, System Architecture, Application Architecture, Component Architecture. This is not necessarily focused on software design, or systems analysis, or organisational information strategy.
|
derek.munneke |
Latest page update: made by derek.munneke
, Oct 4 2006, 1:18 AM EDT
(about this update
About This Update
No content added or deleted. - complete history) |
|
Keyword tags:
Architecture
Australian
computer
sig
Society
software
More Info: links to this page
|
| Started By | Thread Subject | Replies | Last Post | ||
|---|---|---|---|---|---|
| spwhite | SaaS: the next religion (part 3 and final) | 0 | Feb 14 2007, 1:19 AM EST by spwhite | ||
|
Thread started: Feb 14 2007, 1:19 AM EST
Watch
Look at how IPIX going out of business caused panoramas on thousands of websites to stop working. What if you used Flickr to share images between 500 employees when Yahoo logins became compulsory? Does that update to the online word processor break your templates? What are the business impacts of any, all, or various combinations of the local or SaaS systems failing? How can you audit the systems of your SaaS providers to pass your own ISO9001 (and other) accreditation?
With your own systems, you plonk one server centre here, and you plonk another one somewhere else. Whap in a bunch of fibres, throw in some failover, and even that simplest of beginnings has certain basic guarantees that are axiomatic. You can verify the backup procedures. You can test upgrades for your combination of packages. So what's my motivation in writing this? I want something more interesting as the next fad. Something good that has a chance of working. Don't bore me with another tragic mistake like Java. Steve. (*) The One True Editor is joe! |
|||||
| spwhite | Saas: the next religion (part 2) | 0 | Feb 14 2007, 1:18 AM EST by spwhite | ||
|
Thread started: Feb 14 2007, 1:18 AM EST
Watch
If you're the customer, you need to consider other factors. What is your company trying to do? Is it a solution for your problems? What are the costs and benefits of making a commitment? The problem with SaaS is that it adds another solution that has to be maintained, without removing the need for existing approaches. And, like outsourcing, it also adds a new frontier of unmanageable problems for the company.
Are you going to upload confidental data to whatever AJAXy Web3.0 server has the best interface? Will your internal emails be stored in a different legal jurisprudence, allowing competitors to launch nuisance suits to access internal records? How will you comply with Sarbanes-Oxley when your IT system is in 8 different countries around the world? What if a botnet attacks a router in Belgium in the middle of your sales presentation? So you still need your servers and networks, the same as before. And now you have trouble moving data between your machines and the 8 different outsourced SaaS services, not to mention the different licenses, service levels, and the brutal fact that you're just another customer to them. What if their software update breaks one of your business processes? How long will it take for you to find out? How long for them to understand your setup and fix theirs? http://www.techcrunch.com/company-index/ While I have no idea what most of the above companies or websites offer, it does illustrate that the chosen services adds up to a unique combination that has a natural tendency to intricately weave itself into how things are done. So when 37signals has its middle-of-the-night maintenance, your company gets knocked out at peak hour. (continued next post) |
|||||
| spwhite | SaaS: the next religion | 0 | Feb 14 2007, 1:16 AM EST by spwhite | ||
|
Thread started: Feb 14 2007, 1:16 AM EST
Watch
Bill asked for a reaction provoking rant, so here's one!
SaaS, the latest religion First up, I'm a hardware sort of guy. I take requirements and convert them to software, servers and networks. I worked in Defence and now run my own company. Much like Xtreme Programming and other religions, an ecosystem of ideas can be self-referential enough to create a consistent worldview. If only other people would do this, everything would work properly! In the longer term, nature cannot be fooled and debates are settled by practical realities. Think of the old battles that have been fought, which are irrelevant these days. The trick is to work out which of the new ideas will stay around like emacs vs vi (*), or disappear like Cobol vs Fortran. If you remember a few years back, outsourcing was all the rage. Now companies are quietly hiring people back, with overseas staff being made into regional offices. Multi-national corporations operate a lot slower than smaller companies, and outsourcing was the equivalent of splitting the local office across several timezones. The idea this time around is to centralise the software instead of the systems. The people who write the software are the people who maintain the servers that run the software. Instead of a local LAN to the server room, the Internet provides the connectivity instead. It sounds like a new approach that could work, so why not! At least, that'd be what you'd say if you're selling SaaS to a customer. In the few years before SaaS becomes another failed silver bullet, it could be a pig's trough for all the consultants and experts and authors and journalists who rake up the buzzword bingo higher and higher until the house of cards falls and the next fad comes along. (continued next post) |
|||||