Saturday, November 11, 2006

Research -- Google

There's something deeper to learn about Google from Gmail than the initial reaction to the product features. Ignore for a moment the observations about Google leapfrogging their competitors with more user value and a new feature or two. Or Google diversifying away from search into other applications; they've been doing that for a while. Or the privacy red herring.
No, the story is about seemingly incremental features that are actually massively expensive for others to match, and the platform that Google is building which makes it cheaper and easier for them to develop and run web-scale applications than anyone else.

An overlooked feature that made Google really cool in the beginning was their snippets. These show a few sample sentences from each web page matching the search.


Consider the insane cost to implement this simple feature. Google has to keep a copy of every web page on the Internet on their servers in order to show the piece of the web page where the search terms hit. Everything is served from RAM, only booted from disk. And they have multiple separate search clusters at their co-locations. This means that Google is currently storing multiple copies of the entire web in RAM. That would really require a lot of memory for the purpose.


Google has taken the last 10 years of systems software research out of university labs, and built their own proprietary, production quality system. It's a distributed computing platform that can manage web-scale datasets on 100,000 node server clusters. It includes a petabyte, distributed, fault tolerant filesystem, distributed RPC code, probably network shared memory and process migration. And a datacenter management system which lets a handful of ops engineers effectively run 100,000 servers. Any of these projects could be the sole focus of a startup.


The most obvious concern for Gmail is storage. They can't lose emails and more importantly, would never be expected to be down. The backup system used isn't the conventional one. RAID isn't any good. If a RAID disk fails, it requires human intervention to replace the disk, and in case more disks fail, it can lead to data loss. Moreover, it requires hot swap trays which are an expensive piece of equippment. RAID has a problem with high end availability of data at the server level.


Google has 100,000 servers. If a disk fails, it is left to be reclaimed/repalced later. Hardware failures need to be instantly routed around by software.


Google has built their own distributed, fault-tolerant, petabyte filesystem, the Google Filesystem. This is ideal for the job. Say GFS replicates user email in three places; if a disk or a server dies, GFS can automatically make a new copy from one of the remaining two. Compress the email for a 3:1 storage win, then store user's email in three locations, and their raw storage need is approximately equivalent to the user's mail size.


The Gmail servers wouldn't be top-heavy with lots of disk. They need the CPU for indexing and page view serving anyway. No fancy RAID card or hot-swap trays, just 1-2 disks per 1U server.
It's straightforward to spreadsheet out the economics of the service, taking into account average storage per user, cost of the servers, and monetization per user per year. Google apparently puts the operational cost of storage at $2 per gigabyte. It is assume the yearly monetized value of a webmail user to be in the $1-10 range.


Google has 100,000 servers.


Any sane ops person would rather go with a fancy $5000 server than a bare $500 motherboard plus disks sitting exposed on a tray. But that's a 10X difference to the cost of a CPU cycle. And this frees up the algorithm designers to invent better stuff.


Without cheap CPU cycles, the coders won't even consider algorithms that the Google guys are deploying. They're just too expensive to run.


Google doesn't deploy bare motherboards on exposed trays anymore; they're on at least the fourth iteration of their cheap hardware platform. Google now has an institutional competence building and maintaining servers that cost a lot less than the servers everyone else is using. And they do it with fewer people.


They must have a little internal factory to deploy servers, and the level of automation needed to run that many boxes. Either network boot or a production line to pre-install disk images. Servers that self-configure on boot to determine their network config and load the latest rev of the software they'll be running. Normal datacenter ops practices don't scale to what Google has.


Google is a company that has built a single very large, custom computer. It's running their own cluster operating system. They make their big computer even bigger and faster each month, while lowering the cost of CPU cycles. It's looking more like a general purpose platform than a cluster optimized for a single application.


While competitors are targeting the individual applications Google has deployed, Google is building a massive, general purpose computing platform for web-scale programming.
This computer is running the world's top search engine, a social networking service, a shopping price comparison engine, a new email service, and a local search/yellow pages engine. What will they do next with the world's biggest computer and most advanced operating system?

Sunday, August 27, 2006

Pluto loses status as a planet

Astronomers have voted to strip Pluto of its status as a planet.

About 2,500 scientists meeting in Prague have adopted historic new guidelines that see the small, distant world demoted to a secondary category. The researchers said Pluto failed to dominate its orbit around the Sun in the same way as the other planets.

The International Astronomical Union's (IAU) decision means textbooks will now have to describe a Solar System with just eight major planetary bodies.

Pluto, which was discovered in 1930 by the American Clyde Tombaugh, will be referred to as a "dwarf planet". There is a recognition that the demotion is likely to upset the public, who have become accustomed to a particular view of the Solar System.

The scientists agreed that for a celestial body to qualify as a planet:
  • it must be in orbit around the Sun
  • it must be large enough that it takes on a nearly round shape
  • it has cleared its orbit of other objects

Pluto was automatically disqualified because its ighly elliptical orbit overlaps with that of Neptune. It will now join a new category of dwarf planets.

The critical blow for Pluto came with the discovery three years ago of an object currently designated 2003 UB313. After being measured with the Hubble Space Telescope, it was shown to be some 3,000km (1,864 miles) in diameter: it is bigger than Pluto.

2003 UB313 will now join Pluto in the dwarf category, along with the biggest asteroid in the Solar System, Ceres.

Named after the god of the underworld in Roman mythology, Pluto orbits the Sun at an average distance of 5.9 billion kilometres (3.7 billion miles) taking 247.9 Earth years to complete a single circuit of the Sun.

Saturday, July 15, 2006

Two moons in the sky

It's going to be a rare sight so remember - don't miss the opportunity
to view 2 moons in the sky -> 27 AUGUST 2003.

This is something interesting.....Not everyone able to sight and view
2 moons in the sky...only come in about every 2-3 generations.

Never again in your lifetime will the Red Planet be so spectacular!

This month and next month the Earth is catching up with Mars, an
encounter that will culminate in the closest approach between the two
planets in recorded history. The next time Mars may come this close is
in 2287.

Due to the way Jupiter's gravity tugs on Mars and perturbs its orbit,
astronomers can only be certain that Mars has not come this close to
Earth in the last 5,000 years but it may be as long as 60,000 years.

The encounter will culminate on August 27th when Mars comes to within
34,649,589 miles and will be (next to the moon) the brightest object in
the night sky. It will attain a magnitude of -2.9 and will appear 25.11
arc seconds wide.

At a modest 75-power magnification Mars will look as large as the full
moon to the naked eye. Mars will be easy to spot. At the beginning of
August, Mars will rise in the east at 10p.m. and reach its azimuth at
about 3 a.m. By the end of August when the two planets are closest, Mars
will rise at nightfall and reach its highest point in the sky at 12:30
a.m. That's pretty convenient when it comes to seeing something that no
human has seen in recorded history.

Saturday, June 10, 2006

The Sensex....down the drain...

BSE index dropped a hefty 700 points in a bliff. The former rise was because of a global trend. I caused oil prices to rise, gold to be more expensive, other metals, silver and aluminium, also got costly. Iron soared the skies.

All this was a result of heavy industrialization all over the planet. In India too, FII's are investing pretty much and setting up industries. They require metal for the builtings and structures. In a fashion, so, all the metals were costly.

But the recent decline got everything down to the ground. The bull is down. Gold is 8000 something from 11000 a couple of weeks ago.

Anything that goes up, must come down. So did the market. Its a law of equillibrium.

Sunday, April 16, 2006

shock

Mom told me to help her with the curtains. They were synthetic. I was in constant touch with them, pulled the rings off of them, took them down. Then as I put my hands under the tap to wash them....
ZAPP!!!!

Electrostatic Shock.

The body lost electrons and attained an unstable negative charge state. It got attrated to the slightly negatively charged Oxygen atoms of water and produced static electricity.
ZAP...hahahahaha

Friday, April 14, 2006

Indo-US Nuclear Deal.

It is a landmark nuclear deal between the United States and India.

India and USA decided to turn a new leaf in the bilateral relationship. Bush administration declared its ambition to achieve full civil nuclear energy cooperation with India. This would require agreement from the U.S. Congress to adjust U.S. laws and policies. Also required is work with friends and allies to adjust international regimes to enable full civil nuclear energy cooperation and trade with India, including but not limited to expeditious consideration of fuel supplies for safeguarded nuclear reactors.

India will have the same responsibilities and practices and acquire the same benefits and advantages of other leading countries with advanced nuclear technologies.

This new treaty would recognize india as a nuclear weapon state, officially. The deal requires the separation of civil and military facilities in a phased manner and the filing of a declaration about its civilian facilities to the International Atomic Energy Agency (I.A.E.A.).

I'd say that th Secretart of State Condollezza Rice would have to do some serious home work convincing the Congressional members to ammend the US law.
Similar negative approaches have been observed in our home also. BJP has criticised the pact. Ironically, it was the B.J.P. that laid the foundations of the emerging U.S.-India strategic partnership.

Vajpayee argued that the Indian govt had surrendered its right to determine what kind of nuclear deterrent it should have in the future based on its own threat perception. He said that it would require huge costs to separate the civil and military nuclear installations. Aslo it would put restrictions on the Indian nuclear research program.

What I feel about this deal is that it would surely put India amonst the globally ordered nuclear states and would make some allies, also it would make sure not to look at India through the prism of non-proliferation.

Thursday, April 13, 2006

Mind..

Its a bit ironic that sometimes its really hard to explain certain things that we claim to possess.
What is Mind?
I don't think there's a single answer that explains the phenomenon of mind.
Is Mind an object? If one day the electrical and chemical processes in the brain are understood completely, would it explain Mind? Would this understanding account for all faculties including intelligence, consciousness, emotion? I think it is a fascinating phenomenon. And its illusive.

There's a question...what do we learn? Its really simple to answer. Just say something in relation to the things we absorbed and contextually explain it in jargons and certain technical terms to convince that we really learnt it.

But....How do we learn?
Its different from how do we retain memory..retention is an elecrical phenomenon resulting from deposition of calcium in brain cells and capacitating the static within.

Why is it true that "practice makes a man perfect"...how is it possible that our thinking, our intelligence, and our way of percieving instantaneous activities evolves. And...why does it evolve?

What is there that is called the mind....that very phenomenon which makes all this possible.

I'd say that mind is a collection of conscious and unconscious processesthat influence our mental and physcal behaviour.
Mind is where things like ego, love, arrogance, care, hatered, hollowness, pain, laughter, sorrow....all generate..

What the hell is it...........not at all physical atleast...

Thursday, April 06, 2006

Sensex: 11,746

Dalal Street climbed to a new closing high.
The unbridled bull run is continuing without any break.

With the Sensex hitting 11,746, the Index soared with a gain of 467 points in 3 goddamned days. Investors may now shrug of concerns about stretched valuations and I believe that now the focus would be on robust fund inflows and expectations for strong fiscal during the fourth-quarter earnings.

Cement, drug and software were major gainers. The peak was at 11,755 points.

Money is pouring in from the FIIs (Foreign Institutional Investors), and mutual funds.
This year foreign funds allowed a gain of 25% to the index with whoping $4 billion inflow.

The price-earning ratio (P/E) has hit 21.62. A high P/E indicates the high earning potential of the corporate sector. It means that the market is getting overvalued..the market has put a much more optimistic and more expensive value on the shares.

Sunday, March 12, 2006

Age of Altruism

Babies as young as 18 months old can tell when adults need help and will
often do their best to assist.

Lifting things, opening doors, picking up objects.

These early signs of helpfulness is a sign of humsns' altruism.. This is
rare in the animals. Chimps also display such kind of helpfulness. This
suggests that the evolutionary roots of altruism go back to common
ancestors shared beteen humans and chimpanzees.

Tests also show that if an adult struggles to pick up a dropped
clothespin in front of an infant, it quickly crawls over picks up the
clothespin and hands it over to the adult. Chimpanzees performed a
similar task. But human babies were found to perform better at opening
up doors and stacking books in case the adult had his ands full.

Saturday, March 11, 2006

Motion - Gravity - Geometry

Gravity is not due to force, but it is a manifestation of curved space
and time, the curvature being produced by mass-energy content of the
spacetime.

One of the defining features of general relativity is the idea that
gravitational force is replaced by gravitational geometry. The
phenomenon in classical mechanics which were ascribed to the actions of
force of gravity are taken in general relativity to represent inertial
motion in curved spacetime.

Inertial observers, that is the observers in inertial motion, for
example freefalling objects, can accelereate w.r.t each other. For
instance, if we take the case of two freely falling balls, same mass,
same shape, on the opposite sides of the earth, they'd accelerate w.r.t
each other as they approach other.

So their acceleration, or more precisely their motion, would be a
manifestation of the geometry of spacetime in which they exist.