Showing posts with label Stanford. Show all posts
Showing posts with label Stanford. Show all posts

Friday, May 25, 2012

TiEcon 2012

TiEcon 2012, Day 1 (Part 1)

The morning of Friday, May 18th 2012 dawned brighter and earlier for those of us that were part of the organizing committee of TiEcon 2012. It was a phenomenal experience putting the various pieces of the conference together and seeing it grow from discussions around a table to a full fledged conference with attendees from around the world and a speaker list that would rival any other top notch conference. Given that this is an all volunteer effort, I was amazed to see the incredible amount of passion and energy flowing out of people and the attention to detail. Granted that I was doing this for the first time and some of these folks were "veterans" - having done this a few times. Nevertheless, pulling off a conference of this magnitude without a glitch requires utmost dedication and professionalism - which I saw first hand. Even though the hall was not full and there was a long line of latecomers at Registration waiting to get badged, Kiran Malhotra, Vish Mishra, Jai Rawat and Farhat Ali opened the conference on time (which I really appreciated - having been to several other conferences where the start would be delayed so that the hall can get filled). They sprinkled a few quotes in their opening addresses, including this one by Henry Ford, "Coming together is a beginning. Keeping together is progress. Working together is success" - to describe the volunteer effort that helped put the conference together; and this one by a Roman Philosopher, "It is not because things are difficult that we do not dare, it is because we do not dare that things are difficult" - to describe the prevailing spirit of entrepreneurship in the room.


By the time Vishal Sikka was introduced, the keynote hall was not only full, but overflowing. Vishal Sikka's keynote was truly inspiring. The theme for his speech was, "Innovation without Disruption". He began by saying that he wanted to make five observations on the road ahead - his musings based on his experiences of 20+ years at Stanford, Start-ups and SAP.


1. What's Going On


Digital Artifacts are replacing physical artifacts. We are seeing this in music, movies and books. This content ties into our lives in more consumable ways. This results in being able to reach the masses more conveniently. In comparison, Thomas Paine's Commonsense Manifesto that was published 240+ years ago (thanks to the revolution of the Gutenberg Press), reached two-third the population, within a year, and helped seed the birth of America. The connectedness that is accomplished with the digital artifacts creates value and dissolves layers of inefficiency and indirection. He identified the need to understand and evolve complex systems - especially enterprise software - "which is quite primitive," he said.


2. My journey


He identified his journey with three institutions, hosts, paradigms - Stanford, Start ups and SAP. Stanford taught him to learn how to think in an unstructured environment - without boundaries and guidance. He applied this when he was asked by Hasso Plattner to help SAP intellectually renew itself. He treated SAP like a 40-year-old startup while instrumenting this turn-around and did so without  using "processes" or "governance frameworks" - as putting these in place would defeat the purpose of what he was asked to do.


3. Three principles


He used three principles to renew customer landscape without disruption: understanding timeless software constructed around content (application's content, UI, etc.), containers (the runtime) and change (ongoing evolution of the content and the container); systems thinking and evolving systems. These allowed SAP to renew the customer landscape without disruption. The example he provided was SAP HANA that helped simplify enterprise computing and dramatically improve performance thanks to advances in hardware, large amounts of main-memory and massively parallel processing of information. Their 100 TB, 4000 x86 core system is the largest in-memory DataBase system in the world (10x the size of IBM's Watson).


4. Purposeful Work


He urged the audience to take on purposeful projects. Don't waste your time doing things that don't add value. He gave examples of SAP's AppHaus - specifically the RecallGenie iOS app. Another example he gave was Project Charitra - with 10,000+ people here, it is a market for publishing needs and things that one is willing to do.


5. Work Hard. Follow your instinct, your passion and have fun


Vishal quoted the lyrics (by famed poet Gulzar) of a song from a Hindi movie, Kaminey, "Jo bhi soye hai khabron mein, unko jagana nahin" (this is in the song, "Dhan Te Dan". You can listen to the song here - the phrase is at ~ 3:45 mark in the song). The loosely translated meaning of this phrase is - "Don't chase fads. Do your own thing". There are lots of huge, purposeful problems that need to be solved:

  • Bring basic payments, banking to 100s of millions of people
  • Kidney dialysis is being done by ghastly means. Surely, in this modern age there are betters ways to do this
At the conclusion of Vishal's talk, Robert Guest, Business Editor at the Economist had a conversation with him. He asked Vishal about his experiences while growing up in India and a few other "soft ball" questions that he had prepared before hand - not necessarily from the talk that Vishal had just delivered.


Friday, November 11, 2011

Panel on Software Defined Networks

TiE Silicon Valley organized a panel on the next wave in networking - Software Defined Networks (SDN) on Thursday, November 10th. This was well attended (with an audience of over 150) and was easily one of the best panels that I have been to (TiE or otherwise).

Networking has not evolved much in the last 20+ years. Since the first packet switched network in the late 1960s, companies have basically been improving the same architecture and basically improving line speeds and port densities. All that changed with the PhD dissertation of Martin Casado at Stanford in 2007. His thesis was based on the Ethane project, that allows one to build a secure enterprise network with a powerful, yet simple management model. This led to OpenFlow and SDN. Suddenly, there is innovation in networking and there is a lot of excitement around the plain old network!

The first speaker was Stanford's Guru Parulkar. He defined the problem - how networks are vertically integrated, complex, closed and proprietary. By separating the data plane and control plane and having a protocol between the OS and the forwarding devices, it is possible to build network control management applications. He waxed eloquent about the talk given by Verizon's Stu Elby at the Open Networking Summit on Service Provider Networks.

Dan Wendlandt of Nicira spoke next. Dan provided his perspective on SDN - how it is possible to describe high level policies that the system "compiles" down to individual devices. So, a VM can migrate all through the data center and have the security policies follow it.

The next speaker was Anshul Sadana, VP Customer Engineering at Arista Networks. Anshul spoke about OpenStack. He also talked about how virtualization sets networking back one step and how Arista's VMTracer bridges the gap between physical and virtual networks, thus simplifying network operations.

Stuart Bailey of Infoblox was the next speaker. He likened the value (in networking) moving from custom hardware to software to the end of the mainframe era. We are, indeed, at a huge inflection point in the evolution of networking. Infoblox solves the problem of managing pervasive networks.

The final speaker was the host of the evening - Allwyn Sequeira of VMWare. He talked about the trends driving the evolution of the DataCenter:
  • New Application architectures - stateless and scaling out
  • Lower Hardware costs
  • Virtualization of all Hardware components (by 2014, an average server will run 320 VMs)
There is a need to enable dynamic, workload aware networks that has
  • Forwarding plane
  • Control plane
  • Management plane - which abstracts the network and presents the abstraction to the upper layer for logical networks
Overall, I thoroughly enjoyed the session. Each speaker on the panel was an expert and has deep knowledge in their subject matter.