1) When evaluating Drupal, one of the first things companies and evaluators will ask is: What high traffic sites use Drupal?
There are many reasons for asking such a question (SHOULD I LIST SOME? OR IS IT OBVIOUS); Yet there is no such listing or easy answer, and most Drupal sites are low traffic - This is a major road block to the adoption and usage of Drupal to serve high traffic sites.
Usage information can't easily be tracked
- There are no known install base statistics. Even tracking downloads is difficult, let alone installs or usage.
- Some download stats are available for Drupal.org [1] but there are many places to download Drupal [2]
- A conservative guestimate: hundreds of thousands of sites actively run Drupal
- Drupal can be setup to ping a centralized repository [3] but few enable this feature, and the listing info isn't currently public. [4]
- There are a few sites that list, categorize, or attempt to track traffic to Drupal sites - all are based on user submissions, and or editorial choice, they cover only a very small fraction of the Drupal sites on the net:
- Buytaert.net lists high traffic sites and high profile brands using Drupal (~50 sites as of 9/2007)
- DrupalSites is a more open-ended list that categorizes and allows users to rank sites (~1,600 sites as of 9/2007)
- Top.DrupalSites.net tracked traffic via "powered by Drupal" badges they supplied, (has be down since 12/2006).
- Various consulting and services companies, many well known, some not so well know (and there is no real 'top' listing of service companies afik) list their clients and portfolio projects; great, but first you need to penetrate the community and discover them, then you need to go through all their sites… and while someone who REALLY wants to know drupal should do this it takes a LOT of time and effort, and still won't give the full picture
- Spidering the net and fingerprinting (i.e. a CMS survey similar to NetCraft) would be the most accurate methodology - this would take serious computing resources and time, and still has drawbacks [5]
- Anecdotal evidence suggests…
- The vast majority of Drupal sites are low traffic and are likely on shared hosting, a minority are medium traffic, and relatively few are high traffic [6]
- The number of Drupal (core) downloads has been dramatically increasing with each new major release [1], as are the number of businesses that use it [7].
- You can (and an increasing number of individuals and corporations do) run successful, fast-growing, high traffic sites using Drupal.
Many Things To Many People
Drupal is used to power: blogs, personal sites, communities, portals, directories, intranets, extranets, corporate sites, political sites, non-profits, e-commerce, media sharing, independent and professional news/media, entertainment, international or multi-lingual / localized sites, porn and adult sites, etc. (and any combination thereof). If you want to or do run a web site, you should consider using Drupal…
2) This first question often leads to the following realization and follow up question, "Drupal appears to do so many things, for so many people… How can it meet everyone's needs well? And can it meet the needs of a high traffic site like mine?"
the various different constituents have their own concerns and needs
Low Traffic Sites (The Majority)
Low traffic sites range the spectrum of uses listed above, but they likely have the following common characteristics / needs:
- require a high degree of backwards or legacy compatibility due to shared hosting environments
- lack of stack control - can't choose what software is installed, let alone what version, or even tune the software, etc.
- poor architecture - db is likely on the same box, can’t be tuned, etc.
- resource contention due to shared resources/other sites on the same box
- performance at the Drupal level of the stack is typically a non-issue
- in the rare event you get a lot of traffic (you get dugg, stumbled, slashdotted, etc.), you are likely to run into resource issues (bandwidth, i/o, cpu, db, etc.), or have tuning/configuration related problems and will get cut off by your provider LONG before you run into problems with Drupal itself.
High Traffic Sites (The Minority)
High traffic sites will likely fall into one of two categories, mass hosting farms (many individual or federated sites) or one single site. Mass hosting farms, and federated sites run many Drupal installs across a single architecture (or similar/related architectures). The mass hosting tends to be of independent sites, whereas the federations are of interrelated sites. The difference between these and individual high traffic sites is scaling a collection of low, moderate, and even high trafficked sites vs. scaling one single high traffic site.
These 3 categories share the following common characteristics / needs:
- complete stack control with custom and often sophisticated architectures often having requirements like high availability and redundancy
- require custom tuning and configuration of hardware and software (at all layers)
- performance at the Drupal layer IS a concern / issue.
Hosting Farms (which are not the focus of this site) have their own specific needs, some of which overlap nicely with a single site
- A farm likely requires the re-allocation of sites (move them across servers), transfer of data (db or otherwise) across systems, duplication of sites, application of changes across multiple sites (upgrades, patches, settings, etc.) and so on.
- Many of these needs / requirements apply to large traffic sites that employee sandboxes or multiple development environments for a large development team, code is migrated and tested across development, staging, and production sites - much like data and settings might need to be migrated and or applied across sites, and so on…
