RAAK

We all know that the Twitter Firehose is massive. Not massive like a mountain, but massive like a Tsunami.

Exactly how big is it though?

I mean, we get all these figures with strings of zeroes behind them, but what does it actually mean? How can we make this more conceivable?

We had a look at the Twitter Firehose, and equated its sheer size in common terms. And which metric can be more common than SMS? In the process we’ve discovered something quite astonishing .

Twitter, right now, transports more than ten times the amount of messages than the average mobile operator transports SMS’s.

And this number is rising. Rapidly. No wonder we see a fail whale every now and then!

To see something even more amazing, read on …

Twitter

Twitter currently gets an average of 155 million tweets per day. This number is constantly on the rise, and any attempt at consuming it should be made in a very, very scalable manner. Let’s see what kind of bandwidth we’re talking about.

155 million tweets per day is 6 458 333 tweets per hour, 107 639 tweets per minute, and 1794 tweets per second. Each tweet holds on average about 2500 bytes of data. Each byte of data is 8 bits. That means each tweet holds 20 000 bits, or 20 kbit of data.

So, the total throughput of the current firehose is 36 million bits per second.

That is 4 average UK broadband connections (8Mb/s) operating at full blast, 24/7.

SMS

Surely this can’t be much, compared to SMS, can it?

In 2010 6.1 trillion SMS’s were sent worldwide. This is 17 billion messages per day, 696 million messages per hour, 12 million messages per minute, and 193 430 messages per second.

If we take into account that each SMS is only 1120 bits big though, we get a global throughput of 217 million bits per second. This is only six times the volume of the current Twitter Firehose.

Quite surprising, isn’t it?

metric	Twitter	Global SMS
Messages per Second	1 794	193 430
Megabits per Second	36	217

The picture per operator

Now, Twitter is a single entity, and global SMS data is being dealt with by a great amount of mobile operators. 1470, to be precise.

To get a comparative number for a single operator we divide the total amount of messages and throughput by the amount of mobile operators worldwide. That gives us 132 messages per second, and 147 619 bits per second.

That is one 54th of an average 8Mb/s UK broadband connection.

metric	Twitter	Global SMS
Messages per Second	1 794	132
Megabits per Second	36	0.15

And they get a LOT of money for it!

The Future

When will the Twitter Firehose surpass the total global SMS volume? Let’s try and deduct it from a few stats from Twitter’s blog.

This is what we know:

• During the first 1 156 days (roughly), Twitter handled a billion tweets. That is an average of 800 000 tweets per day.
• A year ago, Twitter averaged 50 million tweets per day
• During the month of March of this year, Twitter averaged 140 million tweets per day
• Right now, according to Gnip, Twitter is averaging 155 million tweets per day

Let’s plot that. assuming the year-on-year growth stays the same.

(click to enlarge)

So, assuming that SMS data will not grow at all, Twitter data will surpass SMS data by July 2012. Not too far off then!

And how about the amount of tweets against the amount of SMS messages?

(click to enlarge)

This postpones the crossover date by two years. So, by July 2014, Twitter will handle the same amount of tweets as the global SMS volume today. That is quite impressive.

Now, SMS volume will grow, and Twitter might not continue to grow exponentially, which makes these projections quite optimistic. Still, taking into account that 77% of the world population owns a mobile phone, and that Twitter might have only 21 million active users, these numbers are nothing less than amazing.

Let’s revisit this topic in a year’s time, shall we?

UPDATE – June 2, 2011:
It’s not a year later, but a week, and Twitter has just released new numbers, claiming a billion tweets every six days. That is 167 million tweets per day. Their growth might completely outperform our predictions!