Mr. Self, Would you please share the basis of your statement that, "GT.M faster than Cache on VA benchmarks for 600+ concurrent users on VMS."
Regards, Bain At 03:43 PM 2/6/2003 -0800, you wrote: >Bruce Slater wrote: >>EpiCare is written in MUMPS, but in large installations they run on >>something called CACHE - which is better for large scale applications. > >Please explain why Cache' is better. Having worked with MUMPS for over 20 years and >with >Open Source and GT.M for the last several, I have several reasons to believe that >Cache' >offers a decidedly inferior alternative in cost and performance and scaleability, >especially in the longterm. > > - Cache license $1000 per user process for 100+ users plus $200/user per year. > > - Open Source versions of MUMPS - GT.M and MUMPS_V1 (and others less developed). > > - GT.M has ACID transactions and designed for 24x7x365 continuous operation. > > - GT.M faster than Cache on VA benchmarks for 600+ concurrent users on VMS > (GT.M on VMS is not currently Open Source, but much less expensive than Cache) > > - Premier "Cache" site Brigham and Womens 6000+ user distributed processing system > still running Datatree MUMPS on DOS despite years of effort attempting to switch > over to Cache'. (Has this changed yet? I haven't checked in a couple of years.) > > - GT.M distributed database is Open Source. > - OMI (Open MUMPS Interconnect) protocol between GT.M Linux server and Datatree > clients is significantly faster and more stable in my experience (medium scale) > in VMTH than proprietary Datatree to Datatree DCP (Distributed Cache Protocol). > - PHP client opens up GT.M data to non-MUMPS web applications development. > > - GT.M dynamically compiles MUMPS to native code executables. > Integrates well with host OS. > > - Open Source EsiObjects provides object layer to MUMPS. > > - VistA runs on GT.M > >--------------------------------------- >Jim Self >Chief Systems Developer and Manager >VMTH Computer Services, UC Davis >(http://www.vmth.ucdavis.edu/us/jaself)
