Hernan Question: Web Scaling
Locked
Hernan Question: Web Scaling
 
|
Hi guys
Can you share your experience with Hernan on this question: "In a web kind of architecture, I have doubts about pharo scalability. I don't have information about how well pharo would scale as "web server", you know that it is not the same as a desktop app. I would like to see some info about that, comparing with other techs like ruby, python, php, java, etc. “ Stef _______________________________________________ Pharo-business mailing list [hidden email] http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org |
Re: Hernan Question: Web Scaling
|
|
See the 'In Production' section at the end of this presentation:
http://zn.stfx.eu/zn/pharo-http-fundamentals.pdf My POV is that scaling Pharo is no different than scaling some other dynamic object languages. When you have horizontal scaling, you can do a lot. Getting there is a challenge (sessions, persistence), but that is true for all technology stacks, it is never free (try going beyond a single database). Step one is to be so successful that you need scaling, then things should be good because you have revenue and can worry about getting faster. It never hurts to do a proof of concept and measure, of course. On 13 Mar 2014, at 15:05, Stéphane Ducasse <[hidden email]> wrote: > Hi guys > > Can you share your experience with Hernan on this question: > > "In a web kind of architecture, I have doubts about pharo scalability. I don't have information about how well pharo would scale as "web server", you know that it is not the same as a desktop app. I would like to see some info about that, comparing with other techs like ruby, python, php, java, etc. “ > > Stef > _______________________________________________ > Pharo-business mailing list > [hidden email] > http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org _______________________________________________ Pharo-business mailing list [hidden email] http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org |
|
|
In reply to this post by Stéphane Ducasse
On Thu, Mar 13, 2014 at 03:05:15PM +0100, St?phane Ducasse wrote:
> Can you share your experience with Hernan on this question: > > "In a web kind of architecture, I have doubts about pharo scalability. I > don't have information about how well pharo would scale as "web server", > you know that it is not the same as a desktop app. I would like to see > some info about that, comparing with other techs like ruby, python, php, > java, etc. ? My personal experience is only with running a low volume blog powered by Pharo Smalltalk. However I've just (re-)blogged about the experience of Wanelo, a huge online retail ecommerce site. Many of the points mentioned in that talk should be applicable to web applications built with Smalltalk. http://samadhiweb.com/blog/2014.03.29.scaling.web.postgresql.html My view, reconfirmed by reading those slides, is that a scalable web application isn't about any specific OS, programming language, web framework, ORM/database, etc. Rather it is the combination of all such things, driven by the people involved, backed by the collective know-how available on the net. Here, RoR, to mention the web framework used by Wanelo, has a leg up on Smalltalk due to its rich ecosystem. I will say that Pharo needs better PostgreSQL drivers. I'm concerned that the current PostgresV2 and PostgresV3 native Smalltalk drivers may be susceptible to Bobby Tables. If we can crowdsource adding support for prepared statements in PostgresV3, I'm prepared to chip in. Cheers. -- Pierce Ng http://samadhiweb.com/blog/ _______________________________________________ Pharo-business mailing list [hidden email] http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org |
Re: Hernan Question: Web Scaling
|
|
On Sat, Mar 29, 2014 at 3:56 AM, Pierce Ng <[hidden email]> wrote I will say that Pharo needs better PostgreSQL drivers. I'm concerned that +1 My PostgreSQL/ DBMS driver background is probably not good enough to implement it myself in a secure way but I'm prepared to pay as well if someone has the capabilities to do it.
_______________________________________________ Pharo-business mailing list [hidden email] http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org |
Re: Hernan Question: Web Scaling
|
|
On 30 Mar 2014, at 21:06, Francois Stephany <[hidden email]> wrote:
We can set up a bounty. Could you write down a description? Stef _______________________________________________ Pharo-business mailing list [hidden email] http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org |
Re: Hernan Question: Web Scaling
|
|
I will. On Tue, Apr 1, 2014 at 10:36 AM, Stéphane Ducasse <[hidden email]> wrote:
_______________________________________________ Pharo-business mailing list [hidden email] http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org |
Re: Hernan Question: Web Scaling
|
|
In reply to this post by Pierce Ng-3
Hi Pierce,
On 29 Mar 2014, at 03:56, Pierce Ng <[hidden email]> wrote: > My personal experience is only with running a low volume blog powered by > Pharo Smalltalk. > > However I've just (re-)blogged about the experience of Wanelo, a huge > online retail ecommerce site. Many of the points mentioned in that talk > should be applicable to web applications built with Smalltalk. > > http://samadhiweb.com/blog/2014.03.29.scaling.web.postgresql.html > > My view, reconfirmed by reading those slides, is that a scalable web > application isn't about any specific OS, programming language, web > framework, ORM/database, etc. Rather it is the combination of all such > things, driven by the people involved, backed by the collective know-how > available on the net. Here, RoR, to mention the web framework used by > Wanelo, has a leg up on Smalltalk due to its rich ecosystem. Interesting presentation. I agree with your points. This has been what I felt too lately: to succeed on some scale, really understanding what you are doing (and having a good team) are what matters most and it is something that is independent of most technical details of the technology stack. And that is good news because it means you can use Pharo just as well ;-) > I will say that Pharo needs better PostgreSQL drivers. I'm concerned that > the current PostgresV2 and PostgresV3 native Smalltalk drivers may be > susceptible to Bobby Tables. If we can crowdsource adding support for > prepared statements in PostgresV3, I'm prepared to chip in. To be vulnerable to SQL injection attacks you have to make implementation mistakes at different levels, starting at you web application or web service layer. I would not say that a driver is responsible for that. If you use Seaside for example, or do proper REST, I can't see how you could be vulnerable. Like the point you made earlier: you have to know and understand what you are doing. Regards, Sven _______________________________________________ Pharo-business mailing list [hidden email] http://lists.pharo.org/mailman/listinfo/pharo-business_lists.pharo.org |
«
Return to Pharo Business
|
1 view|%1 views
| Free forum by Nabble | Edit this page |