Copacetic

PutPlace.com is launching real soon now and of course like everybody else on the planet we rolled our own user account management (UAM) system (you know, registration, login, logout, forgot password, change username etc. etc.). Needless to say we used a Framework (in our case Django) but why do I need to reinvent this wheel?

We get our storage from Amazon (and soon our Grid). Our O/S is a flavour of Linux,  our Database is Postgres, we use surveymonkey for surveys and mailchimp (what is it with the simian metaphor?) for mail campaigns, why can’t I pay somebody on a per user basis to manage my user accounts?

What would this service look like? Well pretty much like a credit card payment interface looks like on the web today but with the following features.

•  a bunch of REST endpoints for all the standard UAM functions
  • Register
  • Login
  • Logout
  • Enable User
  • Disable User
  • Set Session Timeout
  • etc.
• Support for session management
• Standard reporting (who, when, from where, how long)
• Support for OpenID
• Java, Python, Ruby and .NET APIs (Rails and Django integration for extra points)
• Global deployment with appropriate SLAs
• Deployment pages with a proper branding kit so people can get up and running quickly
• Ability to extract all my users in .csv format or via an API

I would pay for this service in chunks of hundreds to 1000s of users with all the usual discounting policy for prepayment or purchasing in bulk and the first 50 users are free.

One more pain point on the road to release removed.

I don’t want a free one with strings attached (e.g. LiveID) and I don’t want some non-profit consortium (e.g. the Liberty Alliance). I want to pay  for it and own my data.

Basically a Web 2.0 oriented OnDemand business.

Andrew McNeile gave us the three ingredients for a successful elevator pitch as part of the last Hot House training session for HotHouse 10. He pointed out that an elevator pitch should be,

• Compelling (I have to have this product now!)
• Unique (I can’t get it anywhere else)
• Demonstrate real end user benefit (This solves a real problem for me)

Ciaran Bradley CEO of Sentry Wireless won our internal competition on the day hands down with a brilliant description of his new product KidSafe. 

His website give the full low down on the product better than I could,

Kidsafe benefits for parents

• Prevents mobile bullying and anonymous text messages
• Denies access to premium inappropriate services
• Allows phone to be locked during school or after midnight

Kidsafe benefits for mobile operators

• Differentiated services for parents
• No impact on core network
• Opportunity for increased ARPU /retention
• Device independent
• Works on voice, data and text
• Allows targeting of new markets responsibly

If ISPs (e.g. GoDaddy, Network Solutions) supported a REST API to DNS then applications like Google For Domains wouldn’t have to throw users to the wolves when it comes to configuring CNames and aliases.

Instead you could just give your credentials to Google and they could update your DNS entries directly. Most ISPs discourage direct fiddling with users DNS entries (for good reasons), but with the ability to customise your URL now a feature of Google for Domains, the need for non-skilled users to access this data is becoming more important.

We’ve quietly put up a proper website for PutPlace.com in the fast few days and as it hasn’t fallen over, consider this a mild invitation to run over there and pre-register for the beta. You can do the survey to boot and help make the world a safer place for Digital Content.

What does PutPlace do? Helps you to find, organise, secure and share  that huge and growing pile of photos, video, music, emails, documents and blog content that is building up day by day on you PC, phone, laptop and Media Centre.

So run along over there and register and we’ll send you a private beta invite real soon now.

BTW: Some of you may have come across us by our previous name Secantus, same product different name. It happens, we’ve got over it, you should too

I’ve used Google’s new Custom Search Engine feature to create a first pass at a Web 2.0 Search Engine for Irish companies. Any can contribute to make it better, so please do.

Its also in the sidebar of this blog.

Amazon continues to move away from book selling. After S3 and EC2 along comes fulfilment, They will store and ship your stuff. Great for power sellers.

So now you have unbounded programmatic access to computing and storage resources via EC2 and S3. Anybody who has a moderate amount of programming skill (skilled in the art as the patent people say) and has access to a credit card can slap together their own grid in no time at all. More importantly you can resell a grid to a vertical market.

Now a whole plethora of companies have been trying to flog Open Source stacks to corporate America, and to my mind corporate America (and corporate everywhere else) is not ready to take these companies offerings on board for a number of reasons,

• Certification is no good without delta change detection. Your stack might be good the day it deploys, but is it good after 6 months of production tinkering?
• Most customers need some kind of customised stack, rather than an off the shelf offering, so configurations proliferate and the costs for both the vendor and the customer escalates
• None of these guys wants to play nice with the rest, so how do you integrate stacks from different vendors
• Isn’t Open Source supposed to be free as in beer

However the whole game changes if instead of offering an installable package you offer a managed EC2 image with all the good juice pre-canned and ready to go, and you manage the customers application deployment on top of that image and then cut a new precanned EC2 image with all the good stuff in one bundle. Tinker away all you want and deltas can easily be identified by comparing the production image with a clean install of the original.

Customisation becomes delta management with everything being a managed delta of a base kit and integration of different stacks can be easily tested because the base infrastructure is already deployed into an image ready for merging.

Expect to see lots of people offering layered services on EC2 fairly rapidly with simple Apache, MySQL and J2EE and ESB instances appearing in short order and more complex three tier and N-tier packages  following on rapidly. Also expect the monitoring and management vendors to offer layer packages that plug straight into these environments.

Can Google trump this?

I’m just catching up on my blogs and the one thing that rocks my world is news of Amazon’s new web service EC2. They’ve jumped the gun on Sun and Google to provide a global grid facility to anybody with a credit card. USD$ 0.10 (10 cents) an hour gets you,

… the equivalent of a system with a 1.7Ghz Xeon CPU, 1.75GB of RAM, 160GB of local disk, and 250Mb/s of network bandwidth

You get prebuilt Linux instances to layer your own images on and free interconnect to your S3 storage.  We already use S3 so this is going to increase the security of our application, improve scalability and make it cheaper all in one go!

Total cost for a years worth of compute power 0.10 * 24 * 365 = USD$ 876. Moving stuff from out current hosted environment to our S3 storage was going to cost us a fortune but this could have a radical impact on our most significant cost, bandwidth, where’s that spreadsheet…..

I’ve dealt some digs out to the Cape in the past, but this report from Silicon Republic seems to indicate some serious sales growth.

According to the latest Gartner Dataquest report on application integration and middleware (AIM), Cape Clear holds 18.9pc of the north American ESB market and has experienced revenue growth numbers of 86.7pc. Worldwide, Cape Clear captured 15.9pc of the market.

Congratulations, guys, well done!

Michi Henning writes about The Rise and Fall of CORBA on the ACM Queue website. In the article he describes the history and development of the CORBA standard and details why it has been relegated to a infrastructure backwater. They key reason for CORBA’s failure that resonates with me is the following statement,

The OMG does not require a reference implementation for a specification to be adopted. This practice opens the door to castle-in-the-air specifications. On several occasions the OMG has published standards that turned out to be partly or wholly unimplementable because of serious technical flaws. In other cases, specifications that could be implemented were pragmatically unusable because they imposed unacceptable runtime overhead. Naturally, repeated incidents of this sort are embarassing and do little to boost customer confidence. A requirement for a reference implementation would have forced submitters to implement their proposals and would have avoided many such incidents.