It also allows you to scale up without having to renegotiate contracts
Employed by the big enterprise software vendors, these software police reserve the right to sniff through your infrastructure looking for orphaned instances. And if they find one you may find a big bill landing on your doorstep.
Beware the licensing trap
As Martin Thompson, founder of the ITAM Forum, told Computer Weekly last year, “audits are a tried and tested way to increase revenue” for the enterprise software vendors.
Customers pay dearly for those software licenses. You’re generally looking at a perpetual license fee, along with a software update agreement. The latter adds more than a fifth of the overall purchase price in the case of Oracle’s software licenses, which begin at $47,000 per processor for bare metal installs. It also depends on what kind of processor you use, with x86 and Power incurring different pricing structures in Oracle’s pricing structure.
These licenses are also inflexible, warns Wendy Neu, Head of Americas Database Specialist SA at Amazon Web Services (AWS). Because the terms are so rigid, customers have to pay what amounts to a capacity planning tax. Those Neu talks to complain of two negative effects. First, they can’t easily scale up. That’s a problem in environments which experience volatile workloads. The flipside is that they can’t scale down, leaving them hemorrhaging budget for idle capacity.
“Customers have to buy at a scale that is way above what their normal workload is,” she warns. “They have to buy for peak. And then when they exceed that peak, they have to rebuy the hardware, and renegotiate the licenses with the vendors.”
One alternative is to decamp to the cloud, where you can scale up and down as you see fit – technically, at least. However, software vendors often make it more difficult to run a database instance there.
For some customers, porting an existing legacy database workload into the cloud might carry too many problems with it. Instead, it might be easier to ditch on-premises licensing altogether and buy into a managed service where the license is baked in and you don’t have to deal with the software vendor at all.
Drop the licensing fees
Clients are increasingly looking for this kind of relationship, according to Neu, who says they’re gun-shy after years of draconian fees.
“They don’t want contracts and tie-ins, just like you don’t want cell phone contracts you can’t get out of,” she says. “They don’t want to pay for something that they’re not using or that they’ll need more of and can’t get without intervention.”
Customers who were using AWS-managed databases noticed this flexibility during the pandemic, she explains. Those with home delivery business models suddenly saw orders spike overnight. Without cloud scaling, their databases would have come apart at the seams.
“The customers who were already in the cloud were able to scale up,” she adds. “They didn’t have to ask anybody, they didn’t have to have additional contracts, they just had to spin up a different instance.”
The flexibility that Neu describes goes beyond scaling. It also makes data more portable between different models, she explains.
House of Brick, an IT services management company, argues that one of the biggest challenges for companies running Oracle databases in-house is organizational. The rigid nature of many Oracle database implementations creates silos in which the teams responsible for maintaining them are loath to change anything, leaving businesses unable to evolve as their data platform stagnates.
In one client organization, the situation got so bad that some parts of the company were using Oracle versions that had fallen out of support a decade prior. Developing a cloud strategy for your databases can make changes more tractable, breaking down organizational silos and ensuring that the database team can meet the needs of the business, the services firm said.
More portability
Moving to a cloud-based environment where the stack is managed for you creates a more flexible environment that is more supportive of application changes. It also offers more options for developers. – Companies wanting to optimize their data models are likely to need broader technology solutions than those supported by a relational model. Those might include AI, complex relationship mapping, and perhaps even streaming analytics if they’re working with specialist applications such as industrial sensors. That typically means embracing other data models such as graph, streaming, and document-based systems. They may need to upgrade from rigid tables to key-value stores for better scalability.
Applications might need different data models at the back end. A company might migrate a lot of its data to Amazon’s home-grown Aurora system, for example, while shifting other workloads to DynamoDB’s key-value store for better performance.
Running these new data engines on-premises will be a daunting proposition for many companies, especially if they must combine those models to underpin different applications. If they also wish to upgrade their application architecture to be more modular with container-based microservices or serverless applications, then they’ll need tight integration with their database engine.
This all points to a managed database migration in the cloud where much of the integration is baked into the platform. That will reduce developer effort and learning time, accelerating the time to market.
Companies can use home-baked databases developed directly by the cloud service providers, or if they already have on-premises versions of the various open-source databases, they can move to compatible APIs in the cloud, again eliminating direct maintenance and support contracts.
Challenges to migration
Moving from their old on-premises systems for the licensing and scalability benefits is a good time to explore new technology opportunities and could free you from other architectural constraints. Neu points to the Amazon Aurora relational database, which features read replicas for high availability and performance scaling.
This all sounds utopian, but many companies will face challenges moving from on-premises database architectures built over years and encompassing business rules layered directly into systems by long-departed DBAs. Migration is often expensive and stressful, which is how database incumbents often keep people trapped in inequitable on-premises contracts.
Although it might be like ripping off a gargantuan sticking plaster, migration makes it easier to eliminate what House of Brick calls the ‘Oracle Change Fear Factor’ that exists in on-premises installations. Testing and applying the regular patches that the vendor issues for its database products is unnerving because they sometimes alter operational performance. In a managed database environment, these issues evaporate.
Neu argues that migration is a one-time cost and that the AWS data migration service will help to shoulder the load. This offers a gaggle of tools for assessing source and target workloads and data so that companies can help decide what to move and refactor.
The numbers seem attractive. IDC studied the migration costs, mapping them against the cost benefits, which included not just licensing savings but also the elimination of burdensome, inefficient routine tasks. The customers it talked to reported a 264 percent ROI over three years and a 39 percent lower operational cost over the same period after moving database deployments to RDS.
Companies shouldn’t make the move from an on-premises licensing structure without assessing migration costs, their application roadmap, the projected operational costs of cloud architecture, and the current pain points that they’re feeling in their software licensing relationship. For many, though, an initial evaluation will show that this at least warrants further attention.
In addition to the AWS Database Migration Service tool, AWS offers programs such as end-to-end migration, including assessment, incentives, and professionals to help enterprise customers migrate their critical workloads to the AWS platform. As the company says, this should “really take the migration fear factor out.”