It seems just like yesterday when we got true freedom from the tyranny of a 100mbps cable connected to our laptop and went wireless. Today we have wireless speeds better than what our data center and applications used for a lifetime of computing in the industry with 100mbps in the past.
Fast forward to today,enter, SDN and Openflow. With SDN now we have an architecture for computer networking which is more in tune whats needed today to make the concepts and realities of public and private Cloud based networks possible and make networking more agile and programmable, hence achieving manageability and capabilities to provision faster.
![]()
SDN in essence makes the control plane and data plane separate and allows quick experimenting, provisioning and optimizations of switching and routing policies and provides access with Openflow (which is the communication protocol) to the inner workings of switches and routers that were closed and proprietary. Openflow is the enabler protocol which makes SDN possible.
![]()
I applaud the what companies like Nicira and Big Switch and Insieme are doing in the networking environment with switching and SDN and making our life's easier,data more accessible and networks more manageable and leading the charge to provide stepping stones to our journey to the Clouds.
All that is well and good, but can we please refocus and address the elephant in the room which is storage.
The EMCs and Netapps of the world are not architecturally built to address the exponentially growing demands of the storage industry without draining the customer's pockets and leaving them bleeding.
With Bigdata and the growing needs of storage I think the architecture of storage as a whole needs to be looked at in depth again. Lucky for the industry companies like Google, Facebook and Amazon have very effectively shown us that this can be done, but not by going the route EMC or Netapp would have us go, but by holistically looking at scale out models and by converging compute and storage into a single tier. This would address multiple pain points in the storage industry today and would allow us to stand on the cusp of the next big change in the industry and get us ready to ride the wave of growing our storage needs safely and efficiently.
In my humble opinion, the exponential growth and the need to manage performance and effectively grow storage capacity and infrastructure is the more pressing and crucial problem to tackle.
Enter, SDS, Software defined Storage. The requirements of SDS are the same as SDN, agility,performance and manageability, but this time from a storage perspective, to make it ready for the challenges of Bigdata and the cloud architectures.
![]()
By converging the compute and storage tiers and bringing the intelligence of managing the storage closer to the OS/hypervisors and also by allowing a high performance and a single pane of manageability to both the compute and storage tier's, Nutanix is the only company I think understands and is efficiently addressing this crucial problem with its SOCS architecture, and its ability to grow storage and compute requirements to infinity, hence providing performance, manageability and "pay as you grow" model for the compute and storage infrastructures.
Fast forward to today,enter, SDN and Openflow. With SDN now we have an architecture for computer networking which is more in tune whats needed today to make the concepts and realities of public and private Cloud based networks possible and make networking more agile and programmable, hence achieving manageability and capabilities to provision faster.
SDN in essence makes the control plane and data plane separate and allows quick experimenting, provisioning and optimizations of switching and routing policies and provides access with Openflow (which is the communication protocol) to the inner workings of switches and routers that were closed and proprietary. Openflow is the enabler protocol which makes SDN possible.
I applaud the what companies like Nicira and Big Switch and Insieme are doing in the networking environment with switching and SDN and making our life's easier,data more accessible and networks more manageable and leading the charge to provide stepping stones to our journey to the Clouds.
All that is well and good, but can we please refocus and address the elephant in the room which is storage.
The EMCs and Netapps of the world are not architecturally built to address the exponentially growing demands of the storage industry without draining the customer's pockets and leaving them bleeding.
With Bigdata and the growing needs of storage I think the architecture of storage as a whole needs to be looked at in depth again. Lucky for the industry companies like Google, Facebook and Amazon have very effectively shown us that this can be done, but not by going the route EMC or Netapp would have us go, but by holistically looking at scale out models and by converging compute and storage into a single tier. This would address multiple pain points in the storage industry today and would allow us to stand on the cusp of the next big change in the industry and get us ready to ride the wave of growing our storage needs safely and efficiently.
In my humble opinion, the exponential growth and the need to manage performance and effectively grow storage capacity and infrastructure is the more pressing and crucial problem to tackle.
Enter, SDS, Software defined Storage. The requirements of SDS are the same as SDN, agility,performance and manageability, but this time from a storage perspective, to make it ready for the challenges of Bigdata and the cloud architectures.
By converging the compute and storage tiers and bringing the intelligence of managing the storage closer to the OS/hypervisors and also by allowing a high performance and a single pane of manageability to both the compute and storage tier's, Nutanix is the only company I think understands and is efficiently addressing this crucial problem with its SOCS architecture, and its ability to grow storage and compute requirements to infinity, hence providing performance, manageability and "pay as you grow" model for the compute and storage infrastructures.