Hardware Platform Updates – Haswell Refresh and Configure to Order Support for NX-3000 Series

By Prabu Rambadran
| min

We are excited to announce a couple of big changes to Nutanix’s line of hardware platforms. Starting March 15th 2015, we will support Intel’s Haswell processors. The change will be rolled out over the next couple of months across our entire hardware portfolio. Haswell is the successor to Ivy Bridge microarchitecture, which is the de facto processor in all our hardware platforms today. You can read more about Haswell here.

Why is this meaningful? Enterprises IT shops refresh storage every 3-4 years. Typically, the storage refresh cycle is independent of server refresh as arrays are shipped by different vendors and managed by a different IT teams. If there are advancements in Intel microprocessors, graphic chipsets etc., server vendors are quick to refresh their servers with the latest and greatest hardware components, enabling customers to buy best-in-class technology. An enterprise-grade storage array has at least a couple of storage controllers, which are x86 servers that use the same processors as servers. Since the storage refresh happens less frequently, and support for newer hardware architecture is completely reliant on the storage vendor, customers are typically ‘stuck’ with older hardware components till the storage vendor supports these architectures and the customer refreshes the storage array.

Hyperconverged systems let you refresh storage at the rate of servers. Since compute and storage are shipped as a single appliance, and customers can grow their deployment as needed, IT managers can now add new hyperconverged platforms with Haswell processors to their existing clusters, taking advantage of the latest technology advancements.

So what does this really mean to you?

A simple comparison of performance of Ivy Bridge vs Haswell on a transactional test workload is shown below.

By just moving from the previous generation Intel platform to the latest one, you instantly get a performance boost for your applications, amongst other things. In contrast, if you were to rely on the storage vendor to test, validate and support the latest processor architectures, you will be skipping at least 2-3 processor refresh cycles. With Nutanix, you now have the opportunity to interject processor advancements every 12-18 months, as opposed to once every 3-4 years.

This brings about a profound shift in how you procure infrastructure – instead of having to wait for your vendor to support a transformational hardware technology and stall your business, you buy infrastructure when you need them. As hardware gets refreshed, you add these new nodes to the same cluster and load balance depending on what the Apps need, without any downtime. The conversation has now shifted to what you really want as opposed to what the vendor wants you to do.

The second announcement that we want to make is support for Configure To Order (CTO) for our 3000 platform series. With this, you have the ability to right size your hardware platform to exactly meet the need of your workloads by choosing the right amount of compute, memory and storage that your Apps need. This is the second platform for which we are enabling CTO, after the NX 8000 series and over the next few months our entire hardware portfolio will move to the CTO model as well.

The new NX-3060-G4 platform will be available starting March 15th. The table below lists different hardware components that you can configure. You can also go to the hardware platforms page to learn more about this.

So what does this new platform offer that the previous NX-3000 doesn’t?

A few things – with the 12-core option VM density will increase by over 20% and you can now pack a lot more VMs in a single node than ever before. With support for 1.2 TB and 1.6 TB SSDs along with 2TB HDDs, you now get more storage across both the hot and cold tiers. With the option to add a NIC providing 4X10 GbE ports, you not only get additional network bandwidth, but also the ability to isolate network traffic and assign them to specific ports. For example, the Controller VM traffic can be assigned to a specific port, VM traffic that needs to be secure and isolated can now be assigned to a dedicated port and so on.

One of the true benefits of being a 100% software-defined solution is that we can roll out new platforms that take advantage of the latest hardware updates without any delay or updates to our software. At the end of the day it is about enabling customers to run whatever workload they have on a platform that they most prefer and for software to figure everything else out. Glad that we are able to do this at Nutanix.