Show HN: SpreadServe: QuantLib Calcs, Quandl Data, Serverized XLS with XLL and RTD

http://54.148.111.119:8888/ycb_quandl_pub.xls/Bootstrapping

by osullivj • 11 years ago

3 7 11 years ago

54.148.111.119

osullivj4• 11 years ago

Designed for automated operation of pricing, risk and financial modelling spreadsheets on a server. Supports XLLs, RTD & VBA. WebSocket & RDB integrations so you can take those spreadsheets off trader desktops and share one centralized golden copy of trader developed numerical logic with all users, with no custom coding. The link shows SpreadServe running on an AWS host. Sign up for the beta at spreadserve.com if you want to play with your own copy.

mikes25• 11 years ago

Can the user upload a spreadsheet and run it? Or is the setup/bootstrap more complicated? Is there any custom coding required to get it working?

osullivj4• 11 years ago

No custom coding required. Server side config is needed to setup any XLL the spreadsheet uses; drop the .xll on the server localFS, and add one line of plain text to a config file. Then, if a user is permissioned, they can upload on the repository page, click the 'Run' button next to the SpreadServeEngine on the dashboard page, then load the new sheet.

magmasystems• 11 years ago

I have users who are running into problems with slow performance and crashes when doing larges recalcs with 32-bit Excel on their desktops. Can SpreadServe help?

osullivj4• 11 years ago

Yes: enterprise server class monitoring and logging mean you have a much better chance of nailing the root cause with detailed logging and crash dumps. The scriptability of the SpreadServe environment gives you a better change of reproducing issues. However, SpreadServe can't control the internal implementation details of the 3rd party XLLs it loads. It does provide detailed logging on the behaviour of those XLLs though.

mcqb• 11 years ago

In order to feed the generated prices into our trading systems - what API's do you provide and what format would the data be sent in (JSON, XML etc)?

osullivj4• 11 years ago

Currently SpreadServe has two interfaces for building integrations with other systems: a Python API, and a socket server implemented with Tornado and the Python API. SpreadServe's internal message format is JSON, so to push market data into a pricing spreadsheet you'd code a socket client in Java, C# etc, and send JSON messages. Our RDBMS connector uses this approach, and is coded in Java. If you can code in Python then the Python API exposes more implementation options. We're planning C & Java APIs that will enable integration modules in those languages can avoid talking to the socket server.

Show HN: SpreadServe: QuantLib Calcs, Quandl Data, Serverized XLS with XLL and RTD

Related Stories