Enterprise software development

conrd

Good afternoon all,
We are currently setting up a business and are looking at a make-v-buy decision on our Enterprise software. It is a core part of the business model and requires maximum flexibility.

Our experience is mainly around pre-configured enterprise software (SAP, Oracle, SAGE) all of which would be far to expensive to customise, so we feel that a 'ground-up' build could be the best way to go about things.

The key questions for the moment are;

1.Open source will be critical at the beginning for adequate support, but what should be used....C++, Ruby, .net, how far up stream should we go to get it right.

2.It needs to be able to handle, hi-volume transactions and 'large' graphic files on the platform, which has raised the SQL / noSQL question, but we are very hesitate to use Schemaless data bases for a business platform, (although we don't have direct experience of using it)

3. The user interface, can we bash this out on our own, or do we need to get proffesionals in ( and how to ensure the UX links up to the front-end)


Looking forward to hearing from you all,

Thanks

SalteeDog

What is the function of this software ? Financials, Order Management, Logistics, HR, e-Commerce? All of the above? "Enterprise Software" is a bit vague.

Colonel Panic

Enterprise software is a bit of a catch all term. Is it for accounting, CRM, some sort of asset management, all of the above and more? Have you got a functional specification?

conrd

Agreed, the term is deliberately vague as the platform requirements are not fully fleshed out yet.

Currently the main focus will be order management, capacity modelling, and product lifecycle management and hopefully use a standard finance package at the back-end.

At the moment we want to make an informed decision on how best to build the flows taking into account cost of changes and reliability

stevenmu

Are you planning on doing all of the development work in-house or contracting it out? If you are going to undertake the development in-house, I'd suggest that you get a good architect, or possibly even a couple of them, on board early. They can help guide you through these kinds of choices before you go down the wrong path and get stuck there. If you are contracting it out I'd suggest talking to a couple of companies and getting them to explain their approaches, in particular if you have some functional requirements drawn up they may be able and willing to start estimating costs for you.

conrd wrote: »

Good afternoon all,
We are currently setting up a business and are looking at a make-v-buy decision on our Enterprise software. It is a core part of the business model and requires maximum flexibility.

Our experience is mainly around pre-configured enterprise software (SAP, Oracle, SAGE) all of which would be far to expensive to customise, so we feel that a 'ground-up' build could be the best way to go about things.

Are you really REALLY sure about these things? If you think customising software is expensive have you looked at how expensive it is to develop and maintain from the ground up? A team of developers is an expensive thing to have, and ground-up solutions typically require a lot more man-power than customising commercial-off-the-shelf software. Even ignoring the obvious costs, you also have to think in terms of the time to develop the solution, ground-up solutions can often take a lot longer. And don't forget that you'll need things like testing, QA etc. Actually quality of software should be an important consideration for you too, companies like SAP, Oracle, Microsoft etc can invest hundreds of millions, even billions, in hiring huge teams of the cleverest people out there to spend a few years writing their software, and still they end up with issues, but almost invariably COTS software is more reliable than custom developed.

1.Open source will be critical at the beginning for adequate support, but what should be used....C++, Ruby, .net, how far up stream should we go to get it right.

Again, are you really sure this is needed? I'm only really familiar with Microsoft's stack and not so much with Oracle/SAP or any open source enterprise solutions, but you might be surprised by how much customisation is available. Just as one small example, Microsoft Office is not open source, but is highly extensible through Macros, Extensions and Automation and the Office Suite can be customised to do a huge amount more than you would think, all without touching the source code. Similarly other Microsoft products, like SQL Server, SharePoint, Dynamics CRM, Biztalk etc allow for huge amounts or programmability while still leveraging the functionality that they provide all without touching the source. Microsoft are particularly good for enabling developers to customise and build upon their products, but I'd expect Oracle/SAP/Sage etc to be good too.

I'd also suggest that the underlying technology isn't too important at this stage, as long as it's something reasonably mainstream. Unless your business is software development, then you shouldn't focus too much on what language is used and more on the effectiveness and efficiency of the overall solution.

I should point out that I mostly do ISV work so I could extol it's virtues all day long. My typical project involves taking COTS software like Office, SharePoint, SQL Server, Dynamics CRM, Biztalk etc and spending a few weeks or months tweaking and configuring and customising it to meet a customers needs, and writing the code to glue it all together.

With all that said, sometimes the requirements are just too far away from what a COTS system can deliver and you do need to develop something from the ground up, but if it's possible at all I always try to direct people towards leveraging existing solutions as much as possible.

stevenmu

conrd wrote: »

2.It needs to be able to handle, hi-volume transactions and 'large' graphic files on the platform, which has raised the SQL / noSQL question, but we are very hesitate to use Schemaless data bases for a business platform, (although we don't have direct experience of using it)

SQL Server supports Remote Blob Storage, which allows storing of large files outside the database on some other storage medium (typically a standard file system), while keeping all of the metadata in the database for simple organisation and searching etc. I'm not sure but imagine Oracle and others support similar mechanisms, but there's all kinds of ways of implementing this type functionality either with existing platforms or with a completely custom one.

ChRoMe

conrd wrote: »

1.Open source will be critical at the beginning for adequate support, but what should be used....C++, Ruby, .net, how far up stream should we go to get it right.

Well .net is not open source (well you if are a glutton for punishment you could risk going Mono, but I'd imagine bad things will happen).

Writing this in C++ also just seems like an incredibly bad idea.

Just my 0.02c

Colonel Panic

The entire ASP.Net MVC stack and Entity Framework is open source now, or at least the source is available.

awec

I'm not sure open source is what you want.

To me it sounds like you want something that will be easily supported, basically something that's not so obscure where if something breaks you'll have to hire in an extremely expensive contractor to fix it for you (i.e. SAP).

The .NET stack will have an absolute tonne of support out there. It's probably one of the best supported frameworks available at the moment given it's popularity. You'll also have no problem finding quality talent in this area if it ever gets to that stage of needing more bodies.

If it were me, that's without doubt where I'd be looking. It used to be the case that the Microsoft stack sucked compared to the LAMP offerings but that certainly isn't the case any more.

SalteeDog

Advice above from Stevenmu is good. You don't want to go re-inventing the wheel so customised off the shelf is likely to be the way to go unless what you are doing is incredibly unique. The only elements I would consider developing entirely bespoke are those which may give you a unique competitive advantage or which perhaps constitute part of your product or service. Of course you don't need to go the full-blown SAP or Oracle implementations straight off - look for small or medium sized business equivalents.

amen

Open source will be critical at the beginning for adequate support

You do realise that Open source support is generally one of three types:
A: None
B: Provided by the open source developers but on an adhoc basis
C: paid for and very expensive similar to MS/Oracle/SAP etc

What happens when the open source project developers stop working on the project?

It needs to be able to handle, hi-volume transactions and 'large' graphic files on the platform

At this stage open source or not doesn't really matter. You need someone who knows what they are doing and has experience in writing similar systems.

hopefully use a standard finance package at the back-end

does your selected system have to integrate into your finance package?
Does you Open source project currently integrate ?

You would be better off determining your business requirements and from that figure out what the system has to do and then determine if Open Source or MS/SQL Server, SAP, JAVA/Oracle is the way to go.

Either way you are going to be paying some money.

Currently the main focus will be order management, capacity modelling, and product lifecycle management and hopefully use a standard finance package at the back-end

there is software out there that does this so why re invent the wheel?

It is a core part of the business model and requires maximum flexibility.

Do you have experience in running IT Project and delivering Enterprise projects from the design, specification, testing and implementation ?

No matter solution you choose there will be a lot of work involved.

croo

Adempiere is a very popular ERP written in JEE tech.
It has swing or web (ZK/ajax) user interface.
It's extremely flexible.
It's stable with implementations all over the world.

While it will work out of the box I would see it as a good starting point for companies who really need a custom solution - mostly because OOTB is mostly required by small business but the myriad of configurable options in Adempiere make it potentially complex. A huge amount can be customized with no code changes at all - but if you do need code changes then the application is designed to allow these changes to be easily applied without changing the core system which is important when upgrading! Of course if the enhancement is to the core this is one of the few truly open business app projects with no one company owning it 0 it belongs to a community and is open to all and all are encouraged contribute to enhance and address those weaknesses!

It's multilingual, multi client, multi org, multi warehouse with functionality covering most areas of business.
Have a look at the demo
http://www.adempierecloud.com/webui/
You can login to a demo client called GardenWorld using the userId/pw GardenAdmin/GardenAdmin

It does have some weak areas but as a truly open project people are encouraged contribute to enhance to address those weaknesses!

gargargar

If you are building a large enterprise system you definitely need some help. It is probably unlikely you will get a package that will do what you want so you will probably need to have it developed.

I would echo some of amen's comments:

1) Why make an assertion that it needs to be open source? I presume you need there to be talent pool of people to build and maintain it.

2) There are business applications using NoSQL dbs, just not ones that require ACID transactions.

3) Only you can answer this. If you get it built by someone else then you can work on the look and the wireframes.

I think you really need to get someone who knows their stuff. Especially if this is large scale development. The idea of making the decision on how to store your data, or how you should build the application without really knowing technology, is most likely going to end in tears.

mewso

ChRoMe wrote: »

Well .net is not open source

So do people normally opt for php instead and rewrite it?

awec

mewso wrote: »

So do people normally opt for php instead and rewrite it?

I think the OP was just a bit mixed up around "open source", hence the confusion on the thread.

mewso

awec wrote: »

I think the OP was just a bit mixed up around "open source", hence the confusion on the thread.

Well my point being that most people would assume when open source is mentioned they mean an open source app regardless of language. I don't think the OP was mixed up. That someone would feel the need to point out .net is not open source is what bugs me. Only the truly adventurous would be seeking a language they could improve on. Where I work open source = php which gets on my goat.

Short version - I forgot the sarcasm smiley.

SalteeDog

My reading is that the OP is looking for 'OpenSource' insofar as he is not dependent on a Service provider/vendor (SAP, Oracle..whatever) whenever he wants to add a feature or customise his Enterprise system. It not quite the same as Open Source as we understand it.

amen

My reading is that the OP is looking for 'OpenSource' insofar as he is not dependent on a Service provider/vendor (SAP, Oracle..whatever) whenever he wants to add a feature or customise his Enterprise system.

In that case he better have to access to a team of developers who can make the required changes.

If the software is business critical and a business differentiator/strategic asset then the op should really have his own in house IT team.