BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
The concept of converged infrastructures, while expensive upfront, mitigates heterogeneous data center drawbacks.
Over the years, heterogeneous data centers have curated the highest-performance products for the budget. But mixing systems and software from an array of vendors always poses the potential for integration, management and support problems.
Converged infrastructure (CI) can be disruptive, especially for traditional businesses with extensive pre-existing data center assets and staff. But if the business is willing to invest in the change, converged infrastructure will ease the burden on IT staff and erase traditional silos that limit productivity.
With major advantages over mixed data centers, converged infrastructures might replace hardware and management tool expenditures on this year's IT budget.
Converged infrastructure products do away with the subtle differences in hardware designs, firmware, expansion devices, operating systems and drivers found in multi-sourced data center infrastructures.
In a traditional data center, an older server's processor doesn't support a hypervisor command set like Intel's Virtualization Technology for Directed I/O, for example. Another server's firmware disables the feature by default or omits support for parts of the I/O chain. In the overarching data center server mix, this lack of integration could throw off deployment objectives until the IT team identifies the problem and mitigates it via hardware upgrades or software patches.
Converged infrastructure vendors do all of the integration and testing work upfront. The data center manager chooses from pre-screened server, network and storage components. If the individual components need alterations for full interoperability or fine-tuning for maximum performance, it occurs on the vendor's time, not yours.
Better management and visibility
The differences between heterogeneous server, network and storage components also adversely affect systems management. Ideally, one management tool would identify, monitor and control every piece of equipment in the data center. With hardware from all different vendors, one tool rarely achieves the same level of granular insight and control across every device. In some cases, it won't talk to the device at all. If you're buying multiple management tools for visibility into and control of a heterogeneous data center, it lowers efficiency, adds training time and increases the chance of errors.
Converged infrastructures typically come coupled with management tools that provide full compatibility with all the components. The management layer sees and controls everything at a granular level, so there are no gaps in system control or reporting. This finally achieves the vaunted single pane of glass data center management scenario.
Stronger system maintenance and support
The challenges of heterogeneous interoperability also spill over into vendor maintenance and support. For example, the server vendors may be unable or unwilling to address issues that they believe are caused by the networking vendor's equipment or software. Mutual blame leads to wasted time, unresolved helpdesk tickets and unproductive service calls while the data center team chases multiple vendors for support. Meanwhile, key applications are offline for unpredictable outage periods.
Single-source support is an important justification for converged infrastructure -- the CI vendor provides all of the hardware's maintenance and service. The "one throat to choke" concept is compelling for converged infrastructure adopters. It's important to read and understand any maintenance and service agreements for a clear perspective on problem resolution and escalation paths with the CI vendor.
Lower data center space requirements
Converged infrastructure deployments will likely shrink and bring order to the data center footprint. CI replaces conventional rack servers, network switches and storage arrays sprawl, where new systems were installed for years in any available rack space without detailed planning.
Rather than a mix of conventional boxes and blanking panels taking up racks, converged infrastructures comprise small form factor system hardware, such as blade servers and high-density storage arrays. This high-density approach offers more computing with far less physical space while minimizing airflow containment and cooling efforts.
CI has become popular for remote and space-constrained data centers. Some vendors even claim such significant reductions in physical space requirements that users eliminated entire data center sites.
Platform for private cloud
Heterogeneous data centers also support private cloud deployments. However, the matters of interoperability, management, support and space are only exacerbated by private cloud projects that demand automation, scalability and self-service.
Some vendors tailor converged infrastructure products for private clouds, such as HP's CloudSystem platform or Dell's varied cloud offerings, but all converged infrastructures naturally suit cloud architectures.
Converged infrastructures offer businesses a faster and more efficient pathway to building and running an enterprise-class data center by creating a one-stop-shop for systems and software.