Adopt Cloud Tools Only When Your IT Team Is Ready
As networking hardware gets cheaper, it’s having an effect on how data centers are connecting to each other. As big providers like Amazon, Google and the rest start using SDN and SD-WAN more widely, the hardware costs are coming down. Those hardware costs used to be the biggest consideration when building and provisioning data centers, but that’s not the case anymore. There’s also a lot more capacity possible, now that it isn’t restricted by the hardware. And performance has improved in these provider data centers as a result. Most of the typically tracked network metrics, like latency, loss and capacity, should still be tracked with these modern data centers. A few other tips on taking advantage of cheaper hardware: Know the performance needed by various apps, consider open source SDN software, and remember that SD-WAN is not a fix for latency and other performance issues.
While cloud providers are doing a lot of the heavy lifting for businesses, mainframes also still play a part in modern IT. They aren’t just gathering dust and staying online through electrical tape and sheer IT will—there are vendors adding new features so mainframes can match the always-on nature of this SaaS and cloud app world. One of these features, in particular, allows mainframe coding language COBOL to be subject to the same quick app testing processes as Java and other mainstream code. However, as with many IT changes over the past decade, there’s a big human element here: The mainframe IT teams of yore are retiring and passing the systems along to newer colleagues. They’ll need to do some work around making sure code quality is up to par in mainframe systems.
While we’re talking about on-premises pain, this stat from Gartner is rather chilling: 90% of cloud ERP implementations will fail by this year. That means that since they made this prediction in 2016, many IT teams have spent time and money on tools that didn’t fix the problems they already had. The challenges of cloud ERP, according to Gartner, are the same as the challenges of traditional ERP systems: higher costs, more complexity and inability to integrate with other tools. For many companies, those traditional ERP problems weren’t ever solved in the first place, so moving to cloud doesn’t add much, especially if users are expecting vendors to take care of cloud ERP issues. As with other new technologies, don’t make the leap if your IT team and company aren’t ready.
It’s understandable for IT teams not to be ready for cloud. Cloud architecture is very different from on-premises infrastructure, and as this expert notes, security has become a big challenge when businesses are moving to the cloud. The typical pattern of building secure data centers—scaling vertically with tightly coupled components—is essentially the opposite of the horizontal scale and distributed nature of the cloud. For example, network security appliance deployments don’t work in the cloud like they do on-premises. It’s useful to consider using cloud-native tools and also to start with the problem IT is trying to solve, rather than assuming an old solution or process will automatically work in the cloud.