Data Acceleration: A Game Changer for Converged Infrastructure
October 22nd, 2013 By: Bill Kleyman
Many organizations are seeking to create a more powerful infrastructure capable of doing more with a lot less. The digitization of the business world has forced IT administrators to rethink how they deploy their data centers.
This is what makes the Cisco acquisition of solid state memory specialist Whiptail so interesting. Right now, many high-end cloud shops will argue that there is a gap between what some integrated systems can provide and the performance required to run new types of workloads. For example, business intelligence and Big Data systems need a lot of compute and storage power to process vast amounts of information. In many cases, these resources are pulled from spinning disks or separate solid-state arrays. So why not integrate the entire process together? Why not create systems that can deliver powerful resources directly to the data and the application?
According to the news release from Cisco – “With the acquisition of Whiptail, Cisco is evolving the UCS architecture by integrating data acceleration capability into the compute layer.”
Yes, Cisco just shifted the playing field when it comes to the cloud and providing a unified front around storage, networking and compute. However, the really interesting part is what will happen with the industry once other large hardware makers realize that this is the way to go.
- Less Data Center Real Estate. Virtualization and high-density computing brought very real benefits from data center consolidation. We were simply able to place more users on better multi-tenancy systems. Now we’re taking that entire process to a new level. Already, converged infrastructures are becoming the foundation of cloud-ready systems. Couple that with an flash-based processing power and you’ve got a whole new type of platform. Imagine this entire powerhouse of a system under one unified roof. This isn’t some little platform we’re talking about either. The Whiptail model allows for scaling from one node to up to 30 nodes. From there, it can deliver over four million IOPS and 360 terabytes of raw capacity. That is a truly staggering amount of resources that can now be directly delivered to applications, data and the user.
- Complete Unified Computing. Take all of your computing, storage and resource needs and place them under one unified computing platform. That’s it. You’re done. Manufacturers like Cisco clearly see that the future of the cloud computing and almost all modern technologies will directly revolve around the capabilities of the data center. More systems will placed within a data center model as this hub becomes the home of to the Internet of Everything. Applications, data, workloads, and entire desktops are now being delivered from a public or private data center model. As more users join the cloud and virtualization revolution, there will need to be core, unified, computing resources capable of handling this demand. This means greater levels of networking throughput, more consolidated computing power, and delivering data and applications from solid-state and flash pool resources.
- Creating the Micro-Cloud. With even greater amounts of density and processing power, the converged infrastructure is bound to get even more compact. Soon, it will be very feasible for organizations to deploy micro-cloud environments to extend their infrastructure to branches and remote offices. Furthermore, the with micro-cloud capabilities, IT shops will be able to deliver even greater amounts of content to the end-user. By incorporating caching and better methods of WANOP, a converged infrastructure can help an organization extend their platform to the edge. These smaller systems will be built around sold-state technologies with the incorporation of a massive amount of compute and networking power. This amount of throughput, bandwidth and resource availability will create an even more robust cloud network.
- Big data = Not a Big Problem. One of the big problems around big data and the ability to quantify massive amounts of information was the processing needs around the compute and storage platforms. Now, with solid-state technologies built directly into a converged platform – we’re suddenly delivery quite a few direct IOPS to these big data workloads. Instead of having separate storage arrays for big data engines, the process can be built directly into a converged system. Not only that, big data processing can also be incorporated into the micro-cloud platform. Organizations which are vastly distributed can utilize converged components at the edge to process and quantify critical data to make the right types of business decisions.
- Future Converged Infrastructure. The acquisition of Whiptail means a new type of Cisco UCS technology. It’s going to create a truly powerful platform with very diverse capabilities. This means that other converged infrastructure vendors are going to be examining their designs as well. IBM and HP are also aiming to create the next-generation in converged platform computing while storage vendors like EMC and NetApp continue to blaze their way into the flash and solid-state market. With this much movement around the unified computing platform, don’t be surprised if you see even more technological evolution around platforms like Violin Memory, Nimbus Data, Pure Storage, Kaminario and others. Also, don’t be surprised if some more big purchases or acquisitions are made around this technology as well.
The converged infrastructure model makes sense. We’re able to put more under one roof and utilize fewer resources to deliver powerful platforms. Now, more organizations are finding the power of solid-state technologies to help offload certain types of processes. Here’s what you need to understand:
- Pricing for solid-state systems will continue to become more affordable.
- The reliability of flash and solid-state platforms is only improving.
- Entire cloud platforms and workloads are being designed to run on flash arrays.
- Storage pools and tiered data designs are helping organizations make the most of their storage environment.
Now you’re able to direct data to the appropriate type of storage while maintaining optimal user experience. As the data center model continues to evolve, the converged infrastructure platform will sit square in the middle of it all. Already, we are seeing the modern data center become the home of “everything” that is IT. Moving forward, both private and public data center models will be tasked with supporting even more users – and a lot more data. All of this will translate to the need for even greater resource optimization around the next-generation converged infrastructure platform.
Actually I find converged infrastructure products can take away the enjoyment of combining hardware together. But the conclusion is a more balanced performing environment; one that’s been arranged to meet imperious specifications with low tolerances for failure.
Thanks for the interesting read Bill