34 |
\terms{Design, Reliability, Performance} |
\terms{Design, Reliability, Performance} |
35 |
|
|
36 |
\keywords{versioned hypermedia, dangling links, |
\keywords{versioned hypermedia, dangling links, |
37 |
xanalogical storage, peer-to-peer, |
peer-to-peer, location-independent identifiers} |
|
location-independent identifiers} |
|
38 |
|
|
39 |
|
|
40 |
Introduction |
Introduction |
44 |
which would make them part of the web: |
which would make them part of the web: |
45 |
|
|
46 |
Documents written with OpenOffice or Microsoft Word |
Documents written with OpenOffice or Microsoft Word |
47 |
can nowadays be hyperlinked just like web pages-- |
can nowadays be hyperlinked like web pages-- |
48 |
but nobody does it. XXX |
but users don't do so. URLs of web pages are pasted |
49 |
|
into documents and emails, but links to other |
50 |
|
documents and emails are rarely used. |
51 |
|
|
52 |
|
Such links could be of great benefit to small |
53 |
|
workgroups, for example, to communicate |
54 |
|
connections between documents better. |
55 |
|
|
56 |
.. links needed that don't break when documents are moved: |
.. links needed that don't break when documents are moved: |
57 |
|
|
58 |
|
We believe that a major reason is that such links |
59 |
|
would break when documents are published, moved |
60 |
|
between folders, or sent by email. Changing all the links |
61 |
|
when moving a document is too inconvenient. |
62 |
|
|
63 |
.. using location-independent identifiers for |
.. using location-independent identifiers for |
64 |
non-breaking links: |
non-breaking links: |
65 |
|
|
66 |
|
In order for links not to break, documents would have |
67 |
|
to be identified independently of their location. |
68 |
|
|
69 |
.. non-breaking links seem not globally resolvable: |
.. non-breaking links seem not globally resolvable: |
70 |
|
|
71 |
Several hypermedia systems assume that identifiers either have to |
However, many hypermedia systems |
72 |
|
assume that identifiers either have to |
73 |
say where a document can be found on the network, or they |
say where a document can be found on the network, or they |
74 |
cannot be resolved globally. |
cannot be efficiently resolved on a global scale. |
75 |
URLs, location-dependent identifiers, break when documents are |
URLs, location-dependent identifiers, break when documents are |
76 |
moved. Link services often query only a select set of link |
moved. Link services often query only a select set of link |
77 |
servers, not the whole network [hill94extending-andalso-carr95dls]_. |
servers, not the whole network [hill94extending-andalso-carr95dls]_. |
82 |
|
|
83 |
.. but DHTs can do it: |
.. but DHTs can do it: |
84 |
|
|
85 |
However, recent developments in peer-to-peer systems have |
Recent developments in peer-to-peer systems have |
86 |
rendered this assumption obsolete. Structured overlay networks |
rendered this assumption obsolete. Structured overlay networks |
87 |
[stoica01chord-andalso-ratnasamy01can-andalso-zhao01tapestry-andalso-rowston01pastry-andalso-maymounkov02kademlia-andalso-malkhi02viceroy-andalso-AspnesS2003-andalso-bonsma02swan]_ |
[stoica01chord-andalso-ratnasamy01can-andalso-zhao01tapestry-andalso-rowston01pastry-andalso-maymounkov02kademlia-andalso-malkhi02viceroy-andalso-AspnesS2003-andalso-bonsma02swan]_ |
88 |
allow location-independent identifiers |
allow location-independent identifiers |
90 |
|
|
91 |
.. can be used to implement external links, backlinks &c: |
.. can be used to implement external links, backlinks &c: |
92 |
|
|
93 |
|
DHTs can not only be used to efficiently retrieve documents |
94 |
|
given their location-independent identifier, but also |
95 |
|
to retrieve metadata about the documents. For example, |
96 |
|
all external links to a document can be found using |
97 |
|
a DHT. |
98 |
|
|
99 |
.. brief summary how DHTs work: XXX |
.. brief summary how DHTs work: XXX |
100 |
|
|
101 |
This, we believe, may be the most important result of peer-to-peer |
This, we believe, may be the most important result of peer-to-peer |