BeShort: Uma nova abordagem para encurtamento de URLs

BeShort: Uma nova abordagem para encurtamento de URLs
Pedro P. S. Freitas1 , Wellington J. Dores1 , Fabr´ıcio Benevenuto1,2
1

Departamento de Ciˆencia da Computac¸a˜ o
Universidade Federal de Ouro Preto (UFOP)
Ouro Preto, MG, Brasil
2

Departamento de Ciˆencia da Computac¸a˜ o
Universidade Federal de Minas Gerais (UFMG)
Belo Horizonte, MG, Brasil
{pedropufop, wellingtonjdores, benevenuto}@gmail.com

Abstract. Microblogs like Twitter are social systems designed to allow users to post messages containing no more than 140 characters. With the wide use of short messages on the Web, the use of URL shorteners are increasingly becoming popular. These systems translate a shortened URL into a new URL, typically with few characters, and redirect requests that target the shortened version of the URL to the original long URL. Although extremely efficient, the centralized architecture of such services can introduce delays to users and have been widely used as a way to obfuscate spam, phishing and malware. This paper presents BeShort, a distributed approach for shortening URLs able to avoid such problems. Our approach consists of replacing frequently terms (e.g. “www” and “.com.br”) for UTF-8 characters that are usually not used in URLs. To test BeShort we built a dataset containing 50 million URLs of two popular URL shortening services. Our results show that the BeShort obtains compression rates as efficient as the rates obtained by existing approaches.
Resumo. Microblogs como o Twitter s˜ao sistemas sociais voltados unicamente para a postagem de mensagens com no m´aximo 140 caracteres. Com o grande uso de mensagens curtas na Web, o uso de encurtadores de URLs est´a se tornando cada vez mais comum. Sistemas encurtadores traduzem uma URL em uma nova URL, tipicamente com poucos caracteres, e redirecionam requisic¸o˜ es a` URL encurtada para a URL longa original. Apesar de extremamente eficiente, esses servic¸os podem introduzir atrasos para seus