Keith Townsend believes Dell's moves into hyper-convergence and software-defined networking (SDN) markets requires more testing.
Dell has been quietly transforming into a much different data center company; it has entered several OEM agreements that leverage its engineering capability and its sales channel. Dell has also entered other potentially disruptive markets by separately teaming up with VMware and Nutanix on hyper-converged solutions and aggressively embracing hardware that supports software-defined networking (SDN).
Hyper-converged systems are a source of envy for traditional server manufacturers. Hyper-converged appliances from providers such as SimpliVity and Nutanix primarily compete against more expensive solutions based on NetApp and EMC products. Hyper-converged appliances are commodity x86 hardware with the value-add of SAN software; this translates into high-profit margins for appliance vendors.
Dell announced a two-pronged approach to its hyper-converged strategy. For starters, it's participating in VMware's EVO: RAIL OEM program. EVO: RAIL overlays VMware software atop of OEM server hardware to provide a hyper-converged platform that runs VMware's hypervisor. The second approach is the recently announced Nutanix-based appliance; Nutanix software will run on Dell's XC series appliances. One major difference between the EVO: RAIL appliance and the Nutanix-based XC appliance is the hypervisor options: The Nutanix-based solution allows customers to run Hyper-V, KVM, or vSphere.
Dell also sparked excitement by being one of the first major OEMs to release a white-box switch, which is a network device that an end user can install on third-party management software. The ideal use case is an organization looking to deploy a SDN using completely open standards.
Dell has partnered with Cumulus and Big Switch Networks as operating system (OS) options for its open network devices. Cumulus and Big Switch solutions blur the lines between server and network devices; either solution can be considered a Linux server with a high-density number of network ports. The coupling of Linux and network ports allows for creative applications and services. Automation is a strong use case for an open switch solution. Since the platform runs Linux, network managers can either purchase software from third parties or write software that runs directly on their switch hardware. If desired, it's reasonable to assume that an organization can completely decouple automated network provisioning from its server management platform.
Another set of use cases centers around distributed applications and services. With the ability to run Linux applications on network devices, an organization could build a distributed application load balancer. For example, vendors such as Coho Data have built storage solutions that run code on switches to control the flow of storage data using OpenFlow.
The question is: How well will Dell's support organization execute on its end of the equation? On paper, all of these moves make sense for Dell and customers; however, hyper-convergence and SDN are new markets for Dell. An advantage to commodity server hardware is the huge amount of support data available for quick problem resolution. Without a lot of case history, it's difficult to know what challenges early adopters will face, and if customers are willing to run mission-critical workloads on Dell's version of these solutions.
The bottom line
Support questions aside, Dell understands that it has to evolve into more than a provider of beige boxes. As businesses are looking toward cloud-level features to enable new ideas, enterprise IT managers are expecting more of vendors. Dell has responded with interesting product categories that will need testing over the coming months before it can claim victory.