deletions | additions       diff --git a/Related work.md b/Related work.md  index fe6a913..7fb2acc 100644  --- a/Related work.md  +++ b/Related work.md 
...
# Related Work  - escrow transactions (Neha got a question about these after her talk) Primary "competing systems"  - Strict serializability  - Rococo (OSDI'14)  - Calvin: Pre-ordered locking (SIGMOD'12)  - Spanner: 2PL (OSDI'12)  - HStore: OCC (VLDB'07)  - Serializable  - Lynx (SOSP'13)  ## Phase Reconciliation for Contended In-Memory Transactions  - \cite{Narula:OSDI14}: OSDI'14, Neha Narula, Robert Morris (MIT CSAIL) 
...
- *Questions*  - what happens when *part* of a transaction has to wait until a joined phase (if run during a split phase)? Does the transaction abort? Suspend and resume and commit later?  - What is the performance like if you have an `incr` and `read` in the *same* transaction, on the same record? Didn't see if there were any performance numbers for that, but seems like it would probably interact badly with the phasing.  - escrow transactions (Neha got a question about these after her talk)  ## Extracting More Concurrency from Distributed Transactions  - \cite{Mu:OSDI14} Rococo - two-phase protocol, implemented on distributed transactional db  - https://github.com/msmummy/rococo  - defer pieces of transactions to allow reordering execution to avoid conflicts  - two-phase protocol to reorder staged computations  - decentralized dependency tracking  - all participating nodes receive all observed dependence graphs  - use deterministic order (*why do we need to track dependence graphs?*)  - do offline checking to find dependences and potential cycles  - which pieces are "immediate" and which are "deferrable"  - "merge" immediates with dependent deferrable so you don't accidentally run them too early  - *Benchmarks:* TPC-C  - excellent scaling (near linear)  - drops off with increased contention  ## Enhancing Concurrency in DTM through Commutativity   - \cite{Kim:EuroPar13}: EuroPar'13, Junwhan Kim, Roberto Palmieri, Binoy Ravindran