Most eNatis technicians quit

eNatis in it's various forms has been in development a lot longer than people realize. The idea of a points system (natis) started in the 80s. It would be an eye opener if we saw how much tax payers money has been squandered. The resignations were probably early retirements. (unless Julius is moving into IT)
 
For starters, why is it a Win32 forms application? This means that each and every point of access will have to cost a few thousand in hardware and license fees. The thing is an online application anyway so why not build a web app instead ? That way it could have been accessed by thin clients running Linux or something that will have marginal costs (hardware only). This would even allow trafficops on patrol to easily access the system via a mobile portal on a smart phone to check driver credentials.

If they really wanted to keep it simple they could run the backend IIS server on Windows Server and connect to a Suse box that hosts Oracle (if performance is needed) or just another Windows Server running SQL server. Even better have regional web servers for the clients and then connect the regional web servers to a central database via a fat backbone.

They really have gone wrong with eNatis from the offset design phase and got their architecture horribly wrong.
 
Last edited:
It is a Java written application that only works on windows. Ha! go figure. I don't even know why they designed this software front-end. It was not designed with the cashier personnel in mind. Why should a cashier be dealing with a mouse, all that is required is a keyboard, they shouldn't have to waste time by clicking if the sequence is always the same. I support the web front-end idea too.
 
Hmmm, not a java coder but doesn't that defeat the point? Java was supposed to be for multi-platforms , yes/no? A web app would have been a much better solution yes.
 
Abe said:
I seem to remember it being an Oracle and Java solution.
Click to expand...

Yes Oracle 10 Database with license costing $2.4million for 3000 users.
Running in a RAC cluster, initially it was dog slow
I suppose thats the reason why they upgraded to RAC and if I remember 4 nodes I am sure more like 8 nodes.

Architecture overview powerpoint presentation by Tasima - contractors
1.2
National Traffic Information System (eNaTIS)
The eNaTIS is a real-time processing system, allowing online access from more than 1 800 sites
countrywide. The system is configured in a two tier structure, employing a centralised database and
centralised processing by remote workstations. A disaster recovery centre (DRC) is also provided for.
The configuration of the two tiers is currently as follows:
Highest tier – centralised data centre (DC) and disaster recovery centre (DRC) maintaining all data;
and
Second tier – 1 800 remote sites with workstations for users.
The following technology is employed by the eNaTIS:
-
HP-UX 11iV31 UNIX operating system for application and database servers;
SUPPLEMENTARY INFORMATION FOR THE EXPRESSION OF INTEREST
-
Windows operating system for all workstation computers;
-
Transmission Control Procedures/Internet Protocol (TCP/IP) network :wtf:
Maybe use IPX/SPX

Oracle Database 10g management system.
Click to expand...
 
Last edited:
Last edited:
Great - no wonder it sinks - its running on Itanics...

To be fair I have seen what an HP Superdome can do HPux and Oracle 10g correctly installed and performance tuned - but I pretty sure that is not what they have in this case. They would be MUCH better off with a central SQL server DB and .NET 2.0 front-end clients. Although the standards they seem to have followed for the current system would result in one of the messiest and ugliest MS solutions imaginable for perhaps its good they did not go .NET. Heaven forbid it should turn out to be a VB.NET thing :p
 
Lol nothing wrong with VB, C# runs on the same Framework they both have their pro's and con's.
Only good thing about windows platform is that the total cost of ownership is less than most other platforms, bad thing is that everyone thinks they can fix that platform ...... I can fix it..................format/reinstall :(
 
Maintenance on VB.Net solutions is a bit of a nightmare since it tends to lend it self more to an "out of control" type developing techniques. And then there is lazy man's friend for namespaces : using System.VisualBasic :p
Usually when we take over a solution like that we convert it to C# before we do anything else on it. You get to fix a lot of WTF logic that way.

With eNatis I was referring to only run the back-end on Windows since this is a cheaper server platform to maintain (it does cost a bit more to obtain). The end user interface would be web based allowing for cheap machines with free OSes and browsers accessing it. Possibly there might be even more than one web interface with some catering for mobile phones so that trafficops on patrol can use regular mobile phones to check licenses and fines and so on.
 
Nothing wrong with VB.net. I have seen some pretty poorly structured C# code in my time

In the end it is about the brain behind the code. Languages dont make code crap, people do.

As far as I remember seeing (a few years ago) the system was web based. Maybe they changed it for some reason? It was always very poor though and the users complained endlessly. Being involved with the data side of things it was clear they were making some pretty poor design choices there for a start. When you start building applications on a bad data design ....
 
Last edited:
Byrd said:
...As far as I remember seeing (a few years ago) the system was web based. Maybe they changed it for some reason? It was always very poor though and the users complained endlessly. Being involved with the data side of things it was clear they were making some pretty poor design choices there for a start. When you start building applications on a bad data design ....
Click to expand...

How were you involved? How would you describe the design errors?
 
I was not involved and I can describe a few, first being not web based. Secondly I would not go Java at all, Oracle yes - Java not.
 
Rouxenator said:
I was not involved and I can describe a few, first being not web based. Secondly I would not go Java at all, Oracle yes - Java not.
Click to expand...

Are you familliar with all the user requirements? If a keyboard is all the user can have? If the users cant be trained on a new web app? If the requirement was to have the same look and feel as the old system? If ORACLE was required? And the app needed to be cross platform?

Just asking. I dont know all that much. Extremely interested here.
 
Web apps are as cross platform as you can get. You can built your web apps to be complete keyboard driven. You can make them look and feel just like the legacy system they have to replace. With AJAX update panels they will not even know it is a web app. Browsers like Opera (free and cross platform) support a start up kiosk mode so the client can just boot and work.

Oracle is extremely powerful - I have only interfaced with it from a .NET front-end perspective and wrote a few PL/SQL stored procs and logic. To behold the sheer volume of data you can push into it if you use something like host array inserts is spine chilling the first time you experience it. But I doubt that they really need Oracle for eNatis. Sure it has more granular object locking for concurrency control and unheard of performance if it is correctly tuned - but it is expensive as hell. They could use a much cheaper MS back-end with SQL Server Database and .NET front-end which will be available to end users on thin-ish clients running Linux with a kiosk mode browser like Opera. Much cheaper to maintain and license.
 
The app - rewrite was started in 2001 As far as i know. Which of these technologies were mature enough to develop an app of this magnatude?
Apparently the first version was indeed a web app. The users didnt want that and demanded a more interactive client which resemled the character based Ingress app they had before - complete with behaviour, key strokes etc.
Look, we were'nt there to make the decisions and we dont know all the facts. All i know is that the system has some major interest from the international market and is regarded as a leader in the domain.
Why not seculate on their forums about this? maybe someone will have the guts to defend the architecture?
 
We need to know a lot more about the client's requirements before speculating. Important questions such as:

1) Did the client specifically request Java or an open source solution?
2) Did they client specify the database of choice? Perhaps they requested Oracle as a lot of governments do.
3) Is there any devices on the workstations that need to be used?
4) Did the government request and open source OS?

Quite often your architecture is dictated to a large extent by the customer.

I personally am not sure that a .NET / SQL Server solution would handle the types of volumes that they are processing (they did 16146480 transactions in Jan 2010, thats about 750000 transactions on working days). We build government solutions and the sizes of these systems are astronomical, thousands of workstations connecting simultaneously with the servers. Server clusters are the norm. Even MS solutions are not cheap. For .NET, what Application server would you use?
 
Last edited:
If you need scalability then Oracle on the back-end and .NET on the front-end is the way to go. I have been involved with systems that manages and queries call data records on an Oracle database that exceeds 40TB in size.

However, as someone pointed out, back in 2001 there was no .NET so they when with what they knew. At this stage if they had to do it again, with the ability of things like AJAX and .NET maturing to framework 4.0 I am pretty sure this would be an ideal architecture to use for a rewrite.
 
Rouxenator said:
If you need scalability then Oracle on the back-end and .NET on the front-end is the way to go. I have been involved with systems that manages and queries call data records on an Oracle database that exceeds 40TB in size.
Click to expand...

It's not the size of the DB thats the issue but the number of transactions. You would need an application server that supports clustering and that will handle the volume of transactions. We are talking about 1500 transactions per second. Thats where the biggest problem with .NET is. IIS is out of the question, perhaps you can look at BizTalk but the licensing there would be in the same region as Oracle Application Server and it feels like trying to fit the problem to a specific solution, rather then the other way around.

However, as someone pointed out, back in 2001 there was no .NET so they when with what they knew. At this stage if they had to do it again, with the ability of things like AJAX and .NET maturing to framework 4.0 I am pretty sure this would be an ideal architecture to use for a rewrite.
Click to expand...

AJAX would add substantially to the server work load when compared to a Java client. Don't get me wrong, I am no fan of Java but, IMHO, it's the right choice of product for a solution of this magnitude. .NET has yet to prove itself with these sorts of large solutions. Why is MS not using it for things like Hotmail?
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X