Tech Field Day 8 – Day 1: Nutanix

Got off to an early start the first day with all delegates meeting up downstairs drinking massive amounts of coffee to start the day. A quick walk down the road and we were at the Nutanix headquarters. I just love Silicon Valley, you walk down the street and see all kinds of technology companies – start-ups and established alike. The weather can’t be beat either – just beautiful but then again, I am comparing it to the upper midwest that has extreme temperature swings.

Nutanix – Compute. Storage. Complete.

To be honest, I did a little research on their technology and watched some of their animated product demonstrations on the website, but didn’t fully understand what the product did. The big banner in the meeting room did get their message across.

The 2u chassis is based on the iSCSI protocol for connectivity and creates this “virtual storage network” that you simply plug these units into to increase your capacity. The distributed system software enables what they call a converged and scale-out storage layer (SOCS – Scale out COnverged Storage) that is targeted at virtualization. Each “node” makes use of SSD as its primary storage layer that consists of a FusionIO card in each unit.

Now, here comes the interesting part of this product. Every one of these “nodes” run virtual machines and have their own hypervisor!

Key Features:

  • Hardware: 2u “Block” – or mini-cluster of 4 nodes. Each of the 4 nodes includes Dual 6-core Intel Xeon 5640 processors, 48gb  ram (expandable to 192 total), a FusionIO card (ioDrive)  and 5 TB – 10GB/1GB + software.  That means a single 2U Block has 48 cores, 1.3 TB of Fusion-io Memory, 192 gb RAM (expandable to 768 GB) and 20TB of SATA HDDs.  You can add more nodes to expand the cluster as needed.
  • 800mb/sec on the fusionIO card to supplement the SATA drives behind it
  • Scalable from 4 nodes to 50 nodes
  • Separation of storage locations for the VMX files and the VMDK’s are presented straight through to the hypervisor
  • 1100 watts per 2u chassis

They left us with the message that they see a storage “Convergence” and that all operations are contained within these 2u nodes.


Where I see this product really taking hold, is the SMB market since it allows a company to start with a minimal footprint for storage, but allows them to grow to a sizable environment without having to rip and replace. The best part about this growth is that the product becomes more resilient to failure and increases exponentially in performance due to the placement of VMDK’s across the block devices. As advertised, the system has a nice and pretty interface for managing it which they previewed in the demo.

To date, the system does not have SNMP traps built into it, but they said that this will be implemented very soon.




Xangati at VMworld 2011

I had a chance to stop by and talk to a few folks at the Xangati booth at VMworld and was particularly interested in their ability to “replay” past events in a ‘DVR’ type fashion for quick analysis of problematic areas. This is done with their Management Dashboards product line (XMD), which has two parts to it:

  • Xangati VI Dashboard – Tracking for both the physical and virtual components that include application insight for performance. This is done with their Health Engine as described here.
  • Xangati VDI Dashboard – This is the tracking mechanism for the virtual desktop infrastructure side of the environment.

The product is deployed as either a physical device or a virtual appliance in OVF format. This XMD virtual appliance uses what they call the Xangati Flow Summarizer (XFS) that pulls information from ESXi and utilizes NetFlow to monitor information on the physical network.

Steve Rodgers was able to give us a quick rundown of their product in this short video: