Last week we focused on cutting costs on the infrastructure side of IT spending, which generally accounts for a large portion of an IT budget. The next target, which can consume as much or more of the budget, is software. We evaluate the viability and availability of open source business applications, hosted services, open source operating systems, and a new way to think about commissioning custom development work.
1. The Crown Jewels
Businesses depend heavily on their CRM, MRP, ERP, and other three-letter acronym business systems. With one notable exception, Sugar CRM, there are no viable open source offerings for these types of applications. Sugar, in fact, is not truly useful until you enable the Professional features in the for-pay add-ons. Companies literally spend millions of dollars implementing these systems, a large portion of which goes to consulting services. The applications are extremely complex, and difficult to re-invent for the sake of open source.
Open source developers tend away from purpose-specific applications, generally speaking. They want to create something that other developers will use and depend on, so their software will develop a loyal following and grow exponentially. Subconsciously that is probably true, but the real reason is likely a lack of interest–you don’t often find open source developers that are interested in business resource planning systems, nor do you find people willing to take on such huge challenges.
In short, you are not going to switch to some magical free software in this area and see a huge saving. Starting off a saving money article with a few examples of how you cannot save money may seem odd. The next section, SaaS, and the final section on custom development work, should illustrate the point that maybe there is a better way. For that reason, keep thinking about your most expensive software as you read on.
2. Software as a Service
Everything from Google Docs and e-mail to hosted CRM systems can be classified as SaaS. The potential for money savings is great here. With a SaaS model you don’t spend money on servers to host applications, or the time required for the systems administrators to become familiar with the application. The user is the only one that needs to know how to use the application, if we’re talking about stand-alone services that do not require integration work.
The SaaS model gets a lot of attention from people looking to shed their locally-hosted business applications as well. Instead of buying hundreds of thousands (per year) worth of software that does way more than you need, why not pay per-use for a SaaS-based application? Integrating these applications, to speak as generally as possible, is often easier.
Say your CRM needs to feed the ERP system every time a customer PO is received. Figuring out how this is done, to avoid having employees enter the data in two places, requires understanding each applications’ interfaces. If two systems are incompatible, you are stuck (usually buying software that interfaces the two). However, if both systems use industry-standard communication protocols, you have the flexibility to implement whatever is necessary. You also have the flexibility to jump ship should the need arise.
Often, SaaS applications are heavily focused on ensuring interoperability, which can save you money in the long run. You certainly are not going to replace your ERP system that everyone knows how to use overnight, but slowly, as more and more applications move to the cloud, you might find it very easy to do. Agility saves money in the long run, which isn’t easy to sell if you’re only looking at a short-term budget.
3. Open Source Operating Systems
Open source operating systems are easy to sell, however.
Linux ubiquity is here, and even if you pay for support contracts, the total cost of ownership is significantly lower compared to Windows servers. The stability is proven, as is the interoperability. Phasing out Windows servers by migrating services to Linux is a quick way to save large amounts of money.
Another common objection to adoption of Linux is staff skill levels. Controversial as it may be, you have to phase out systems administrators that only know Windows. For example, take your average Linux jockey, who has been exposed to the nuts and bolts and knows how the important protocols work, rather than just how one vendor’s GUI works, and thrust new tasks at her. She will quickly master anything and that is the type of sysadmin you want. This rarely works the other way around, yet the Windows-only administrator generally make the same salaries. The sound business decision is fairly straightforward.
4. Clustering and High Availability
Clustering and high availability are often the most expensive and difficult features to enable on your servers. You need rock-solid reliability, but the cost of application load balancers and operating system licenses at least quadruples the cost of each server. Proprietary operating systems gouge you to enable “enterprise” features, but free operating systems tout their prowess in this area.
Linux vendors, such as Red Hat, make all their clustering software open source. CentOS, the Red Hat clone–literally a clone; they just recompile Red Hat’s source–allows IT to leverage the most advanced developments in operating system clustering, without paying a cent. Application-level load balancing and scaling can also be done for free, using numerous open source tools, without paying anything.
5. Custom Development Projects
The IT software budget often gets consumed by contracted labor, especially during integration projects. Something people rarely talk about is developer lock-in. Much like vendor lock-in, developers make most of their money with follow-up work when the scope of the project changes or an upgrade is necessary.
To drastically cut both initial project costs and lock-in risks, consider starting with an existing open source project. Frequently, you can find software that comes close to meeting your needs, but needs a few changes. The developers you hire will already be familiar with the product, and they don’t start from scratch. In the future, you also have the luxury of hiring other experts with that application, which creates competition and better work output.
Often a development project will consist of integrating two proprietary systems. There likely isn’t any open source software you can leverage. In this case, now is a good time to step back, take in the whole picture, and see if some flexibility could be gained from utilizing some of the savings measures mentioned above.
Last week we focused on cutting costs on the infrastructure side of IT spending, which generally accounts for a large portion of an IT budget. The next target, which can consume as much or more of the budget, is software. We evaluate the viability and availability of open source business applications, hosted services, open source operating systems, and a new way to think about commissioning custom development work.
1. The Crown Jewels
Businesses depend heavily on their CRM, MRP, ERP, and other three-letter acronym business systems. With one notable exception, Sugar CRM, there are no viable open source offerings for these types of applications. Sugar, in fact, is not truly useful until you enable the Professional features in the for-pay add-ons. Companies literally spend millions of dollars implementing these systems, a large portion of which goes to consulting services. The applications are extremely complex, and difficult to re-invent for the sake of open source.
Open source developers tend away from purpose-specific applications, generally speaking. They want to create something that other developers will use and depend on, so their software will develop a loyal following and grow exponentially. Subconsciously that is probably true, but the real reason is likely a lack of interest–you don’t often find open source developers that are interested in business resource planning systems, nor do you find people willing to take on such huge challenges.
In short, you are not going to switch to some magical free software in this area and see a huge saving. Starting off a saving money article with a few examples of how you cannot save money may seem odd. The next section, SaaS, and the final section on custom development work, should illustrate the point that maybe there is a better way. For that reason, keep thinking about your most expensive software as you read on.
2. Software as a Service
Everything from Google Docs and e-mail to hosted CRM systems can be classified as SaaS. The potential for money savings is great here. With a SaaS model you don’t spend money on servers to host applications, or the time required for the systems administrators to become familiar with the application. The user is the only one that needs to know how to use the application, if we’re talking about stand-alone services that do not require integration work.
The SaaS model gets a lot of attention from people looking to shed their locally-hosted business applications as well. Instead of buying hundreds of thousands (per year) worth of software that does way more than you need, why not pay per-use for a SaaS-based application? Integrating these applications, to speak as generally as possible, is often easier.
Say your CRM needs to feed the ERP system every time a customer PO is received. Figuring out how this is done, to avoid having employees enter the data in two places, requires understanding each applications’ interfaces. If two systems are incompatible, you are stuck (usually buying software that interfaces the two). However, if both systems use industry-standard communication protocols, you have the flexibility to implement whatever is necessary. You also have the flexibility to jump ship should the need arise.
Often, SaaS applications are heavily focused on ensuring interoperability, which can save you money in the long run. You certainly are not going to replace your ERP system that everyone knows how to use overnight, but slowly, as more and more applications move to the cloud, you might find it very easy to do. Agility saves money in the long run, which isn’t easy to sell if you’re only looking at a short-term budget.
3. Open Source Operating Systems
Open source operating systems are easy to sell, however.
Linux ubiquity is here, and even if you pay for support contracts, the total cost of ownership is significantly lower compared to Windows servers. The stability is proven, as is the interoperability. Phasing out Windows servers by migrating services to Linux is a quick way to save large amounts of money.
Another common objection to adoption of Linux is staff skill levels. Controversial as it may be, you have to phase out systems administrators that only know Windows. For example, take your average Linux jockey, who has been exposed to the nuts and bolts and knows how the important protocols work, rather than just how one vendor’s GUI works, and thrust new tasks at her. She will quickly master anything and that is the type of sysadmin you want. This rarely works the other way around, yet the Windows-only administrator generally make the same salaries. The sound business decision is fairly straightforward.
4. Clustering and High Availability
Clustering and high availability are often the most expensive and difficult features to enable on your servers. You need rock-solid reliability, but the cost of application load balancers and operating system licenses at least quadruples the cost of each server. Proprietary operating systems gouge you to enable “enterprise” features, but free operating systems tout their prowess in this area.
Linux vendors, such as Red Hat, make all their clustering software open source. CentOS, the Red Hat clone–literally a clone; they just recompile Red Hat’s source–allows IT to leverage the most advanced developments in operating system clustering, without paying a cent. Application-level load balancing and scaling can also be done for free, using numerous open source tools, without paying anything.
5. Custom Development Projects
The IT software budget often gets consumed by contracted labor, especially during integration projects. Something people rarely talk about is developer lock-in. Much like vendor lock-in, developers make most of their money with follow-up work when the scope of the project changes or an upgrade is necessary.
To drastically cut both initial project costs and lock-in risks, consider starting with an existing open source project. Frequently, you can find software that comes close to meeting your needs, but needs a few changes. The developers you hire will already be familiar with the product, and they don’t start from scratch. In the future, you also have the luxury of hiring other experts with that application, which creates competition and better work output.
Often a development project will consist of integrating two proprietary systems. There likely isn’t any open source software you can leverage. In this case, now is a good time to step back, take in the whole picture, and see if some flexibility could be gained from utilizing some of the savings measures mentioned above.
When he’s not writing for Enterprise Networking Planet or riding his motorcycle, Charlie Schluting works as the VP of Strategic Alliances at
the US Division of LINBIT, the creators of DRBD. He also operates OmniTraining.net, and recently finished Network Ninja, a must-read for every network engineer.
