[twitter-dev] Re: Percent of total tweets in GardenHose Stream Access

2010-02-28 Thread M. Edward (Ed) Borasky
That's good to know - if Gardenhose was 10X Spritzer, what I'm trying
to do wouldn't be feasible with Gardenhose because I'd have to buy a
bigger pipe into my "lab". I'd probably run out of patience with CPU
usage in Perl and Ruby, too ;-)

On Feb 28, 8:18 pm, John Kalucki  wrote:
> The gardenhose is very very roughly 3x the default access level (aka
> Spritzer). The algorithm is slightly complicated, and the inputs vary, thus,
> vagueness.
>
> -John Kaluckihttp://twitter.com/jkalucki
> Infrastructure, Twitter Inc.
>
> On Fri, Feb 26, 2010 at 11:05 PM, M. Edward (Ed) Borasky
> wrote:
>
> > I don't know anything about "gardenhose", but I do have some data from
> > "sample". Assuming Twitter's published number of 50 million tweets per day
> > is the full firehose (minus spam, according to the blog post), I've
> > estimated that a typical weekday on "sample" is delivering over 3 percent of
> > the total tweet volume.
>
> > My peak from "sample" this past week was 95,006 tweets for the hour
> > starting "2010-02-24 01:59:59 +". The average JSON tweet is about 1400
> > bytes. That peak represents a bit rate of about 300 K bits per second.
> > --
> > M. Edward (Ed) Borasky
> > borasky-research.net/m-edward-ed-borasky/
>
> > "A mathematician is a device for turning coffee into theorems." ~ Paul
> > Erdos
>
> > Quoting Mark McBride :
>
> >  Hard numbers aren't made public, but it's safe to assume "significantly
> >> more
> >> than spritzer"
>
> >>  ---Mark
>
> >>http://twitter.com/mccv
>
> >> On Thu, Feb 25, 2010 at 10:30 AM, rb  wrote:
>
> >>  Does anybody knows roughly the gardenhose access to streaming API
> >>> provides what % of total tweets.


[twitter-dev] Re: Percent of total tweets in GardenHose Stream Access

2010-03-11 Thread Don Park
Based on my recent #s, Gardenhose is streaming ~3% of Firehose at ~22
tweets-per-second.


Re: [twitter-dev] Re: Percent of total tweets in GardenHose Stream Access

2010-03-01 Thread John Kalucki
Wait a few months. Organic growth will eventually drive Gardenhose to be 10x
today's Spritzer.

-John


On Sun, Feb 28, 2010 at 11:59 PM, M. Edward (Ed) Borasky
wrote:

> That's good to know - if Gardenhose was 10X Spritzer, what I'm trying
> to do wouldn't be feasible with Gardenhose because I'd have to buy a
> bigger pipe into my "lab". I'd probably run out of patience with CPU
> usage in Perl and Ruby, too ;-)
>
> On Feb 28, 8:18 pm, John Kalucki  wrote:
> > The gardenhose is very very roughly 3x the default access level (aka
> > Spritzer). The algorithm is slightly complicated, and the inputs vary,
> thus,
> > vagueness.
> >
> > -John Kaluckihttp://twitter.com/jkalucki
> > Infrastructure, Twitter Inc.
> >
> > On Fri, Feb 26, 2010 at 11:05 PM, M. Edward (Ed) Borasky
> > wrote:
> >
> > > I don't know anything about "gardenhose", but I do have some data from
> > > "sample". Assuming Twitter's published number of 50 million tweets per
> day
> > > is the full firehose (minus spam, according to the blog post), I've
> > > estimated that a typical weekday on "sample" is delivering over 3
> percent of
> > > the total tweet volume.
> >
> > > My peak from "sample" this past week was 95,006 tweets for the hour
> > > starting "2010-02-24 01:59:59 +". The average JSON tweet is about
> 1400
> > > bytes. That peak represents a bit rate of about 300 K bits per second.
> > > --
> > > M. Edward (Ed) Borasky
> > > borasky-research.net/m-edward-ed-borasky/
> >
> > > "A mathematician is a device for turning coffee into theorems." ~ Paul
> > > Erdos
> >
> > > Quoting Mark McBride :
> >
> > >  Hard numbers aren't made public, but it's safe to assume
> "significantly
> > >> more
> > >> than spritzer"
> >
> > >>  ---Mark
> >
> > >>http://twitter.com/mccv
> >
> > >> On Thu, Feb 25, 2010 at 10:30 AM, rb  wrote:
> >
> > >>  Does anybody knows roughly the gardenhose access to streaming API
> > >>> provides what % of total tweets.
>


Re: [twitter-dev] Re: Percent of total tweets in GardenHose Stream Access

2010-03-01 Thread M. Edward (Ed) Borasky
On 03/01/2010 06:17 AM, John Kalucki wrote:
> Wait a few months. Organic growth will eventually drive Gardenhose to be 10x
> today's Spritzer.
> 
> -John

That's the message I'm trying to send my blog readers ;-).

--
M. Edward (Ed) Borasky
http://borasky-research.net/smart-at-znmeb

"A mathematician is a device for turning coffee into theorems." ~ Paul Erdős



Re: [twitter-dev] Re: Percent of total tweets in GardenHose Stream Access

2010-03-11 Thread M. Edward (Ed) Borasky
That sounds low - I'm seeing about 3.4% of all *status IDs* coming out  
of "sample" formerly known as "spritzer". I would expect "gardenhose"  
to be delivering about 3 times that, or about 10% of all status IDs.

--
M. Edward (Ed) Borasky
borasky-research.net/m-edward-ed-borasky/

"A mathematician is a device for turning coffee into theorems." ~ Paul Erd?s


Quoting Don Park :


Based on my recent #s, Gardenhose is streaming ~3% of Firehose at ~22
tweets-per-second.





Re: [twitter-dev] Re: Percent of total tweets in GardenHose Stream Access

2010-03-11 Thread John Kalucki
I've added more detail to the Sample section of the Streaming API wiki.

On Thu, Mar 11, 2010 at 1:45 PM, M. Edward (Ed) Borasky wrote:

> That sounds low - I'm seeing about 3.4% of all *status IDs* coming out of
> "sample" formerly known as "spritzer". I would expect "gardenhose" to be
> delivering about 3 times that, or about 10% of all status IDs.
> --
> M. Edward (Ed) Borasky
> borasky-research.net/m-edward-ed-borasky/
>
> "A mathematician is a device for turning coffee into theorems." ~ Paul
> Erd?s
>
>
>
> Quoting Don Park :
>
>  Based on my recent #s, Gardenhose is streaming ~3% of Firehose at ~22
>> tweets-per-second.
>>
>>
>


Re: [twitter-dev] Re: Percent of total tweets in GardenHose Stream Access

2010-03-11 Thread Don Park
You are right. I got the names mixed up. My numbers are for "spritzer". Doh. - 
Don

On Mar 11, 2010, at 1:45 PM, M. Edward (Ed) Borasky wrote:

> That sounds low - I'm seeing about 3.4% of all *status IDs* coming out of 
> "sample" formerly known as "spritzer". I would expect "gardenhose" to be 
> delivering about 3 times that, or about 10% of all status IDs.
> -- 
> M. Edward (Ed) Borasky
> borasky-research.net/m-edward-ed-borasky/
> 
> "A mathematician is a device for turning coffee into theorems." ~ Paul Erd?s
> 
> 
> Quoting Don Park :
> 
>> Based on my recent #s, Gardenhose is streaming ~3% of Firehose at ~22
>> tweets-per-second.
>> 
>