sakura - Fotolia
Organizations migrate to the cloud to remain competitive. But as interest in the cloud increases, virtualization vendors see an opportunity to maximize profits through means that might not benefit those purchasing their services.
Organizations that switch to cloud-first vendors risk exposing themselves to several drawbacks. The disadvantages of cloud services include vendor lock-in, licensing model changes and reduced cost efficiency. Although in some cases it can be a net win, IT administrators should approach with extreme caution.
When cloud migration makes sense
A cloud-first approach can be a positive scenario for both vendors and organizations. For example, consider how VMware provided Horizon as a value-add service across several cloud platforms. In a typical VMware Horizon hosted environment, VDI usage can be extremely seasonal, especially for organizations that hire call center staff on temporary contracts over busier periods of time.
Organizations that build on-premises environments typically have underutilized hardware and software. But with a hosted per-user VDI environment, hardware and software don't go unused, and there is no hardware to maintain across the estate. With many users and a highly variable usage rate, cloud-based platforms can make business sense.
But large organizations don't just migrate from on premises to the cloud in one step; they complete the transition through a long process that includes specialists, engineers and project planners. Cloud and infrastructure providers can assist with the pre-sale and post-sale resources organizations need to make the leap, thereby, bolstering their consultations and services business.
Vendors can also optimize end-user Opex cloud infrastructure and service subscription costs because they own the whole stack; they know where to save on costs for customers. This is extremely important because organizations are exceptionally sensitive to uncontrolled or unexpected costs.
Cloud subscriptions can help increase cloud consumption because elasticity is a huge selling point.
Organizations also don't have to worry about hardware purchases and issues or hypervisor patching. For example, VMware's cloud platform provides organizations with dedicated hardware nodes to run their stacks.
Trying to achieve elasticity on premises can have long lead times -- purchasing hardware, deploying infrastructure, changing management procedures -- but in the cloud, resources are available in minutes. The hardware and setup are the vendor's responsibility.
This isn't to say that organizations should disregard processes. Rather, when organizations optimize certain processes, scaling out takes seconds. And with the elastic cloud, organizations can consume additional services and roll costs into the monthly operational expenditure.
Disadvantages of a cloud-first agenda
If organizations love one thing, it's stability, especially when it comes to revenue. In the world of virtualization, it's no different. The purchase of on-premises hypervisors, such as vSphere and Hyper-V, through a leasing deal alongside company-owned hardware means organizations can use the product to the max while paying a well-defined and agreed-upon rate.
Traditionally, organizations paid a one-time purchase fee, and then several percent a year in upgrade and maintenance fees for virtualization software. Some vendors are worse than others regarding maintenance fees. But when an organization has thousands of sockets, costs can really add up. Good negotiators get the upgrades at a reduced price, while some sites hold on until the grim end before upgrading to a version.
For virtualization vendors, this purchasing approach can cause revenue streams to be very unpredictable. Plus, if organizations consider an application inferior, they might skip a version altogether, with the vendor deriving no income from that version.
Once vendors convince organizations to adopt the cloud with motives such as flexibility, scalability and removal of hardware purchases and maintenance costs, vendors can:
- Prevent organizations from skipping a software generation. This locks organizations into accepting the cost whether or not they use the product.
- Upsell technology. Consulting and providing technology as a service means revenue growth.
- Make licensing changes. Before the advent of sockets featuring many tens of CPU cores, vendors would license software by socket. That meant a finite amount of compute power per socket. But with multi-core sockets, organizations have 20 times or more performance per socket with no increase in cost. Some vendors, such as Microsoft, switched to a per-core licensing model, but that can still leave money on the table. With a cloud-based model, vendors have more control. And organizations trying to figure out exact cloud costs face challenges. For example, they might not account for sudden spikes in usage or workload growth.
- Lock organizations into the cloud. Once in the cloud, repatriation becomes difficult and expensive. The concept of repatriation from the cloud -- or moving back on premises -- is difficult, but it's gathering momentum. Moving back on premises means technology evaluations and potential changes in infrastructure, which can sometimes be enough of a hassle to keep organizations in the cloud.
- "Optimize" pricing. The vendor has cloud pricing flexibility on an extremely granular level, and customers are often the bargaining chip between software and cloud vendors. For example, Microsoft increased its on-premises prices and the cost to use Microsoft products on rival clouds. This makes alternative vendors more expensive, and it removes software deployments previously permissible on dedicated cloud services. As a result, providers such as AWS, Google and Alibaba saw increased prices for Microsoft software. Use the cloud of choice or pay more is the message. Cloud vendors are reducing the prices of many products, and on-site licencing costs continue to increase significantly each year. Essentially, the more expensive on-site licencing costs become, the more attractive the cloud looks. Either way, it ends up costing organizations more.
It should come as no surprise that major virtualization players are pushing a cloud-first agenda. Virtualization is extremely mature and, in terms of hypervisor functionality, there isn't much left to add. Pair this with a vast amount of organizations going cloud-first in their approach to infrastructure, and it leaves hypervisor vendors with no option but to embrace the cloud. Either that, or lose out to SaaS, IaaS or other cloud providers. Adapt and survive is the message of the day.
Ultimately, this puts organizations at the mercy of the vendor, and escape becomes very difficult. Some vendors are more aggressive than others. For example, several sources have said that Citrix is extremely keen to get users onto its cloud service. However, once in the Citrix cloud, there is no going back to on-premises licencing, and any advantageous licencing deals will be a bitter . The licences get switched to cloud consumption and organizations can't reverse them. Citrix is far from the only vendor doing this; a lot of vendors want to get organizations into the cloud to have predictable revenue streams going forward.
Moving to a cloud-based virtualization vendor can be a net-win approach, but organizations should use extreme caution. Once made, many decisions can't be undone. Organizations must crunch the numbers, talk to their stack vendor and understand the disadvantages of cloud services. Those that effectively can have a good outcome.