131 |
|
|
132 |
Do we want to keep GISP or move to Kademlia or some other? |
Do we want to keep GISP or move to Kademlia or some other? |
133 |
|
|
134 |
|
RESOLVED: |
135 |
|
|
136 |
|
I think we can use GISP as long as we do not deploy Storm |
137 |
|
into a "production use", i.e., we use GISP more like a P2P |
138 |
|
testing platform for Storm rather than "the best and the final" |
139 |
|
choice for Storm. |
140 |
|
|
141 |
|
However, the problem is that how soon the deployment |
142 |
|
will happen (I don't know anything about the milestones :)). |
143 |
|
|
144 |
|
Currently, I see Kademlia as the best alternative to us, but |
145 |
|
as we just got finished a working Storm/GISP |
146 |
|
implementation, do we really want to move an another |
147 |
|
algorithm/protocol immediately ? |
148 |
|
|
149 |
|
OTOH, it's obvious that better and better P2P algorithms |
150 |
|
are invented in the future. The question is that do we |
151 |
|
want to wait those algorithms, or do we want to use |
152 |
|
the algorithms which are currently available ? |
153 |
|
|
154 |
|
ISSUE: |
155 |
|
|
156 |
|
Do we want to implement currently available P2P lookup |
157 |
|
algorithm ourselves, or do we want to develop a novel |
158 |
|
P2P lookup algorithm ourselves ? |
159 |
|
|
160 |
|
|
161 |
Plan |
Plan |
162 |
==== |
==== |
163 |
|
|