Storage architectures are becoming increasingly more complicated. There is more data coming from more applications. There are more users and developers. And there are more operational challenges as IT tries to maximize flash performance across multiple protocols. The storage network is the cardiovascular system providing the organization the data it needs as it continues to modernize itself.
One of the challenges of the modern organization is it is more distributed, with more offices than ever. Many of these offices are large enough that they need small scale data centers to support the location’s users and applications. In addition, many smaller organizations need a high performance storage network much sooner than in the past. These companies are buying all-flash and hybrid arrays and need a network to match the potential performance of these storage systems.
To help the distributed data center and the data center class small business, Cisco is introducing a 32G Fabric Switch. Customers can purchase the switch with as few as 8 ports active and then via licensing activate ports incrementally up to 32. Cisco didn’t cut corners though, it has a complete feature set including full line rate bandwidth, FC-NVMe qualification and deep visibility. It also includes Cisco’s next generation ASIC that provides real time telemetry. (Detailed below).
Larger environments need to support the various protocols and topologies required by the applications they support. They can’t afford to build a separate infrastructure for each. Storage networking convergence is a must. The storage architecture needs to support a variety of speeds and connection types so it can reduce the cost of the overall infrastructure.
To help enterprises, Cisco provides unified ports on its Nexus 9300-FX Platform Series. Any port on these switches can support either Fibre Channel (FC), fibre channel over Ethernet (FCoE) and traditional Ethernet (IP) at the latest speeds (10/25G Ethernet or 16/32G FC). The result should lead to a 50% reduction in the number of cables adapters and ToR switches.
Finally, it is more important than ever for environments large and small need to understand and constantly access the state of their environment. The speed at which IO flows means that telemetry data about the environment needs to be captured in real-time. If the storage architecture status is just sampled, too much can happen between IOs.
The problem has been that in the past implementing a TAP into the networking architecture was the only way to capture real time telemetry data. While these TAPs are passive in operation, they are invasive in implementation. The TAP requirement limited their use to only the largest of data centers and only after a problem occurred. These data centers don’t benefit from proactive data management.
Cisco helps solve this problem with its next generation ASIC that can provide real-time telemetry data without impacting switch performance. It eliminates the need for the TAP. Most interesting is Cisco provides open access to the telemetry data. The openness means the telemetry data can be fed directly to a storage analytics tool like Cisco’s own Data Center Network Manager (DCNM) or Virtual Instruments’ VirtualWisdom.
Virtual Instruments brings over a decade’s worth of expertise in fibre channel SAN performance management. It’s VirtualWisdom product is an analytics tool for the storage network providing a real-time visual view of its status. But it does more than just present data. It will automatically identify weak spots or problem areas and even provide recommended solutions.
The Virtual Instruments’ challenge was, in the past, to get the real-time data it needed to do its job, required TAPs, which limited customer adoption. Now, thanks to Cisco’s telemetry integration, Virtual Instruments is a software only solution, making it much easier for customers to get started. Now, a Cisco infrastructure customer can reap the value of proactive SAN management before a problem ever occurs.
The best IT cost reductions come from driving out inefficiencies. They are both immediate and long lasting. An entry level switch with enterprise class functionality means that organizations can create professional storage architectures sooner rather than later. Creating a switch that, from a unified port, can support all of the organization’s protocols reduces cable and switch costs and simplifies rack design. And providing open telemetry data allows the proactive management of SAN resources, reducing the chances of costly unplanned downtime.