Blobs vs Smart blobs
Hey All, We currently have document storage using the blob data types. This is a legacy architecture used prior to the introduction of smart blobs. The size of the current storage for these blobs is about 10TB. The blobs themselves have traditionally averaged around 80k per blob and we've been running 32k page size. We're starting to see the size of the documents vary more and more as we move to different document types, so that we are now getting documents that can often exceed 10MB and sometimes exceed 1GB. Blobs have always worked really well (we've been using them for 20+ years for this application) but wanted to revisit this issue because of our changing document size issues. What would we gain and lose as a result of moving to smart blobs? Thanks!
I'm not the best person to answer this... But two advantages come to mind: 1- Ability to use replication 2- Ability to spread the LOBs in a table over a set of smatblob spaces (in a round robin fashion) There are probably other programming advantages that I would have to review. Regards. On Fri, Sep 30, 2016 at 4:14 PM, ANTHONY LANDRY <tony@clerk.org> wrote: > Hey All, > We currently have document storage using the blob data types. This is a > legacy > architecture used prior to the introduction of smart blobs. > > The size of the current storage for these blobs is about 10TB. > > The blobs themselves have traditionally averaged around 80k per blob and > we've > been running 32k page size. > > We're starting to see the size of the documents vary more and more as we > move > to different document types, so that we are now getting documents that can > often exceed 10MB and sometimes exceed 1GB. > > Blobs have always worked really well (we've been using them for 20+ years > for > this application) but wanted to revisit this issue because of our changing > document size issues. > > What would we gain and lose as a result of moving to smart blobs? > > Thanks! > > > ************************************************************ > ******************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > -- Fernando Nunes Portugal http://informix-technology.blogspot.com My email works... but I don't check it frequently... --94eb2c048e5a96bc5e053dff92bb
Hello. I would place a bet on NoSQL feature. This will enable you to handle multiple document kinds, regardless of their different sizes. But in 11.50 latest releases, I think there was already the possibility to use XML stored in SBlobs. Check the manual for XML according to your engine version, and use the best that fits your engine. Hope it helps. Best regards. Alexandre Marini Em 3 de out de 2016 9:51 PM, ANTHONY LANDRY <tony@clerk.org> escreveu: Hey All, We currently have document storage using the blob data types. This is a legacy architecture used prior to the introduction of smart blobs. The size of the current storage for these blobs is about 10TB. The blobs themselves have traditionally averaged around 80k per blob and we've been running 32k page size. We're starting to see the size of the documents vary more and more as we move to different document types, so that we are now getting documents that can often exceed 10MB and sometimes exceed 1GB. Blobs have always worked really well (we've been using them for 20+ years for this application) but wanted to revisit this issue because of our changing document size issues. What would we gain and lose as a result of moving to smart blobs? Thanks! ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum.
I tend recommend smartblobs for reasons Fernando mentioned however if you are don't need replication and you have a smaller system (< few hundred GB blobs) I have found blobspace page optimized for the blob size tend to read perform better Cheers Paul Paul Watson Oninit www.oninit.com +1 913 387 7529 On unit® is a Registered Trademark of Oninit LLC On Oct 3, 2016, at 19:59, Fernando Nunes <domusonline@gmail.com> wrote: > I'm not the best person to answer this... But two advantages come to mind: > > 1- Ability to use replication > 2- Ability to spread the LOBs in a table over a set of smatblob spaces (in > a round robin fashion) > > There are probably other programming advantages that I would have to review. > Regards. > >> On Fri, Sep 30, 2016 at 4:14 PM, ANTHONY LANDRY <tony@clerk.org> wrote: >> >> Hey All, >> We currently have document storage using the blob data types. This is a >> legacy >> architecture used prior to the introduction of smart blobs. >> >> The size of the current storage for these blobs is about 10TB. >> >> The blobs themselves have traditionally averaged around 80k per blob and >> we've >> been running 32k page size. >> >> We're starting to see the size of the documents vary more and more as we >> move >> to different document types, so that we are now getting documents that can >> often exceed 10MB and sometimes exceed 1GB. >> >> Blobs have always worked really well (we've been using them for 20+ years >> for >> this application) but wanted to revisit this issue because of our changing >> document size issues. >> >> What would we gain and lose as a result of moving to smart blobs? >> >> Thanks! >> >> >> ************************************************************ >> ******************* >> Forum Note: Use "Reply" to post a response in the discussion forum. > > -- > Fernando Nunes > Portugal > > http://informix-technology.blogspot.com > My email works... but I don't check it frequently... > > --94eb2c048e5a96bc5e053dff92bb > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum.