Posts Tagged ‘TechTalk’

ACM Tech Talk – November 2008 Edition

ACM November 2008 edition had two talks:

Dr Ravi Gudi from Honeywell research gave a talk on Decomposition Paradigms for Large Scale Systems. Dr Ravi Gudi has studied at IIT Bombay and is currently working with the research team at Honeywell Technology Solutions. Using the example of building a chemical refinery and a railway workyard, he explained how to design and manage a large scale system by breaking it down into smaller parts using various techniques. Two of the techniques for decomposition that he talked about are model co-ordination and goal co-ordination. He talked about optimisation problems such as placements of sensors in an area to collect data and transmit it back effectively. The other problem he talked about is communication and feedback while accounting for the delay in transmission. There has been some interesting research in this area at CMU. Some of these fundamental problems see applications in social networks, scheduling in multi-core systems and building Internet scale distributed systems such as CDNs. For example the cascade algorithm invented at CMU can be applied for ranking blogs by number of unique breaking stories. It tells you which blogs should you follow to get the maximum amount of information with the least amount of effort. The same Cascades algorithm can also be used for getting optimal placment of sensors at least cost. Communication in Large scale systems is also facilitated by techniques such as the gossip protocol. Amazon uses the gossip protocol for communication between it’s machines in it’s AWS datacenters. However this communication method overloaded Amazon’s servers and led to a outage.

Prakash Venkatraman, Senior Architect at Oracle talked about using Signature matching and rule-based matching systems for diagnosing faults in systems (root cause analysis). Taking the example of Oracle Support system, he explained how by looking at the error logs and mointoring different parameters, you can build and deduce rules which can be used for fault analysis. Using Bayesian probability and using statistical co-relation techniques such as that used in filtering spam, the whole process of doing fault analysis can be automated to make the work of humans easier.

Google Developer Day, Bangalore, India – October 2008

Google held it’s first developer day in India yesterday at Chancery Pavilion. There was a lot of hype created around the event and lots of people who had registered had failed to get in. I registered early and was lucky to get in. The initial registration process was smooth and Google developer day started on time.

The Keynote:

Dr. Prasad Ram, Google India head, presented the keynote with the active help from Jagjit Singh who demoed various Google products. Dr. Prasad covered three main themes influencing Google’s mission – Client, Cloud and Connectivity.

  • Clients
  • The recently released Chrome Browser and the Google Gears API would come under the client category. Google Gears is being standardised under the web-workers draft tentatively planned to be included in the next version of HTML. (You can follow the updates for the WHATWG on Twitter). Chrome can be used to turn any website or webapp into a desktop shortcut with the help of Google Gears. Google positions this as something new though similar functionality is provided by Microsoft using MSHTA (Microsoft HTML Application), where you can embed presentations, audio, video and interactive features into a nice bundle which can be easily distributed. After this, there was a small demo of the Chart API and Visualisation APIs by Jagjit Singh. Both of these seem interesting and I am planning to use both of these in the near future.

  • Cloud
  • Alongwith Amazon and Yahoo, Google is one of the poster childs of cloud computing services. Dr. Prasad Ram talked about Google’s motivation for moving data and services to the cloud and gave a small demo of Google App Engine (Covered in more detail below). This was followed by a video demo of some of the features of Google Maps. The cool and useful Map Maker application to improve maps was made by a Google team based out of India. I later got a chance to interact with this team in the Google Maps tech track. Though not related to cloud computing, Indic computing, localization and Google Transliteration API was covered under this section. Google is cleverly using crowdsourcing to make it’s maps and voice search better by taking the input from users like us and feeding this data to machine learning algorithms to make both translation and transliteration better. Voice search has been launched in Delhi and Hyderabad.

  • Connectivity
  • Focusing on Clients and Clouds is nice but they are meaningless without connectivity. This brings us to the third theme of connectivity. The next wave of growth is going to come from mobile internet users. No wonder, Google is trying to target this market by releasing Google Android – Google operating system for the devices. Unfortunately there was no session on Google Andriod at this years Google Developer Day. Keeping with this theme, Google has bid for and won the 700 Mhz wireless spectrum auction by the FCC in the past.

The keynote gave a good overview of Google products but overran the allotted time by half an hour due to which the other sessions ran late.

Tech Tracks

I attended the Google Maps API and the Google App Engine tech tracks. The tech tracks were divided into two parts – a tech overview and introduction talk followed by a code lab.

  • Google App Engine
  • Rafe Kaplan – a developer from Google App Engine (GAE) team based in San Francisco – led this tech track. The GAE currently supports only the python runtime and is based on an older modified version of the Django framework. Java support for GAE is also planned. Rafe gave an overview of the different components of GAE such as the datastore, memcache, urlfetch, mail and users API followed by best practices for building a webapp. Later in the code labs, we built a small wiki app using GAE. This workshop could have been better if more details about the backend and internals were included in the talk. The speaker was clearly underprepared for some of the questions which the audience asked him. For example, GAE recently announced limited support for HTTPS and the speaker was not aware of the same. The USB included Python 2.6 runtime and GAE SDK had problems running using it. Fortunately I had Python 2.5 Installed on my machine (being lazy and not having upgraded :) and was able to get the example and make minor changes to it. GAE seemed like a nice platform to play around with but it clearly has some way to go before it can be used for serious applications. This is reflected in the gallery. However, GAE is easy to get started and running and the online docs have lots of information to get you started.

  • Google Maps API
  • The Google India team which built Google Map Maker was at hand for help this talk. Lalitesh Katragadda spoke on Google Maps API. His talk was was extremely informative and had loads of interesting trivia (eg. 80% of the world has not moved beyond the 20 Mile radius around their birthplace). Google Maps is very useful as it is possible to overlay various kinds of information onto the maps. This was followed by a code lab where we built an interesting app – loosely based on Nat Geo Genius. The app is well thought out and illustrates the usage of various API very effectively. The source code and step-by-step procedure for building it is available online at http://maps-gdd2008.appspot.com/ . This excellent code lab was however interrupted continually by WiFi outages which was necessary to read the documentation and see code for the app.

There were several goodies on offer for geeks including a T-shirt and a 2GB USB stick with source code for Windows, Linux and Mac. However I would have enjoyed it more if there were no frequent power and Wi-Fi outages at the venue. Also Google, being a environment-conscious company, should not have unnecessarily wasted paper by printing the detailed qualifications of each speaker on thick paper. Overall it was a good experience and I look forward to attending it next year as well.

ACM Tech Talk – July 2008 Edition

ACM Bangalore Chapter recently had a ACM Tech Talk at CSI Bangalore office at Infantry road. The speaker was Ankita Garg of IBM Linux Technology Centre. Topic of the talk was Linux as an Enterprise real-time operating system .

Ankita started with an introduction of realtime systems differentiating between hard and soft real-time systems. This was followed by discussions on patches applied to Linux to make the kernel pre-emptible, high-resolution timer (which I found particularly useful), syncronisation primitives and threading library. It concluded with a discussion on the current challenges for linux as enterprise realtime system.

It was a very interactive talk and attendees had lots of questions. We ran out of time sadly and Ankita could not give a demo. But if there is sustained interest in operating systems talk then it might be a good idea to start SIGOPS chapter in ACM Bangalore to get all operating systems gearheads to talk together

Get Adobe Flash player