将IIS / ASP.NET站点移植到LAMP？

Question

I run technology for a medium sized company that is about to acquire another medium sized company. Our technology is all LAMP (Linux/Apache/MySQL/PHP), the company we are acquiring is all Microsoft stack (IIS/MSSQL/ASP.NET). None of the developers on staff currently do .NET nor have ever supported Microsoft server infrastructure. I'm having a tough time deciding what to do with the situation...

Do we port all the MS stuff to LAMP (not interested in going the other way for various reasons including my team's personal inexperience with it, the cost of licensing when we are trying to slash overhead, etc)?

Do we run both technologies in parallel with separate teams to support each and write a bunch of middleware so they can talk to each other?

Neither of these choices are optimal. Has anyone ever been faced with a situation like this and how did you proceed? Keep in mind we are talking about large infrastructure in both cases with high traffic volumes and fairly extensive backend systems. Any ideas will be welcomed.

Answer 1

I've never done this before, so take my opinion with a grain of salt. But I would suggest NOT rewriting an existing application. I mean, if it's a 1-page application which just tells you "Hello" when you click a button, then yes, rewrite it in PHP. But business applications that make money aren't as simple as that, and you'll be starting from scratch to rewrite something that took the other company x years to develop. Not to mention you'll have to support and maintain the application you're taking over, even while you rewrite it in PHP.

If you have smart developers on your team now, and they have capacity, they'll be able to learn ASP .NET. But it might be best to hire some ASP .NET resources to help your team learn it and bear the weight (maintenance and support) of the application you're taking over. Your teams can work together to find integration points between the two applications.

Faced with the choice of writing integration points, or writing an entire business application from scratch, I'd take my chances at writing integration points.

Answer 2

As part of the acquisition, are your company taking on the IT support team of the acquisition?

While eventually there are likely to be 'efficiency savings' that they'll want to make from consolidating back office staff, there is a strong argument to keep both teams supporting their 'own' systems in order to keep the lights on.

Then you need to analyse the overlap - do you end up with systems on each stack doing similar things. If so, look to consolidate onto the preferred platform and remove the other. Also look at (regardless of current skills), which stack best needs the business needs in the coming years. LAMP might be perfect right now, but there may be arguments for moving to .net to meet future needs. Then again maybe not, but needs to be assessed.

Is there a business need for the 2 sets of systems to share data? If so, at what level? Creating (web)services to encapsulate shared functionality and make it available to the other system may be one way to go (SOA effectively). Alternately you may need to share a backend initially and have .NET talking to a MySQL databases or somesuch.

Answer 3

This is a very complicated question.

If the two applications provide similar functionality, then I would run both side by side until the one you want to keep has all of the functionality of the other one. Then I'd switch the customers over and eventually throw it away. If the customers are receptive, switch them now.

If they are radically different apps then I'd most likely just maintain both going forward. Given that these are large applications, any rewrite is going to be painful and have a high probability of failure. It's best to just get used to the idea of having different tech stacks in house.

One thing, by maintaining both apps you will be in a better position to keep the acquisition as quiet as possible as far as the client base is concerned. Clients that already use an app typically only change horses if they feel the app they are using is no longer going to be supported. At that point, you can guarantee that some will leave regardless of how good the other system is.

If the acquisition is going to result in a change in marketing (for example, the other company's logo changes etc) then I would again suggest to just maintain both. The clients are going to be nervous enough as it is.

The point of all the above is that this is more of a business problem than a tech issue and boils down to the reasons you acquired the other company in the first place and how you will present it to the existing clients. If the company was acquired for the technology or their client base, then leaving it alone is a good idea.

BTW, I've done this a couple times. The only difference was going the other route from PHP to .Net.

In one case the app was relatively small, but had a huge base of users. We ended up using some URL rewriting rules so that the user base never even knew the app changed underneath them. It was a collection of web services.

In another case, the app was large, had a big user base, and had a very public skin. Again, we heavily leveraged url rewriting to preserve google placement as well as bookmarks. The biggest problem we had was development on the original site couldn't stop while we built the replacement. This presented a lot of challenges in that every feature had to go through both teams. In the end, the project took about 3 times longer than expected but because we had some highly skilled people on it it ultimately succeeded.