Can your business afford to be offline?
The strong market for fault tolerant computing solutions exists for a reason: Some businesses simply can’t afford to be offline. In fact, in the most recent annual study on the cost of downtime conducted by the respected Ponemon Institute, the average cost of downtime for U.S. data centers was $7,900 per minute, with average recovery times of well over one hour. The frequency of outages was as alarming as their cost. Ninety-one percent of the data centers surveyed had experienced an outage within the last 24 months.
In addition to quantifiable costs, system crashes that affect customer-facing applications can do real damage to a brand, resulting not only in a lost sale but a lost customer. The bottom line is that the vast majority of companies with data centers of over 2,500 square feet are looking at significant losses unless they’ve implemented fault tolerance (FT).
FT has been around in one form or another since the mid-1970s, but until recently FT has been a hardware-based solution, and therefore quite expensive. In essence, achieving FT has meant acquiring and operating duplicative hardware at twice the CapEx cost. Also, managing cluster-based FT, one common approach, is complex, and may strain the technical capabilities of smaller IT organizations with limited clustering experience.
Software-Based Fault Tolerance for Multiple vCPUs
VMware has offered a software-based FT alternative for some time in its vSphere product, and that offering has been significantly enhanced in vSphere 6.0. As a result, software-based FT will become a viable option for many more sizes and types of businesses.
The most important advance in vSphere 6.0 from the standpoint of FT is its ability to support multiple virtual CPUs (vCPUs), as opposed to the support for only one in previous versions. This is important because most modern enterprise applications require at least two vCPUs. In fact, a recent VMware customer survey indicated that the one CPU limitation excluded 68 percent of workloads.
Now, multi-processor FT is available for virtual machines (VMs) running as many as four vCPUs, with 64 GB of RAM. This allows administrators to achieve the levels of uptime characteristic of clusters without the need for expensive hardware-based solutions, or the expertise to manage them.
Another FT advantage in vSphere 6.0 is that the primary and secondary VMs don’t need to share the same VMDK. The benefit is that administrators can also achieve fault-tolerant storage with this new approach.
The FT capabilities of vSphere 6.0 will make a big impact on data centers where availability is crucial. Significant use cases include retail point of sales (POS) systems, health care systems, manufacturing applications and financial services. For these cases and many others, FT has never been more cost-effective.
Originally posted on:
A Berkeley-based high tech marketing writer specializing in enterprise solutions such as big data, mobile, security and the cloud.