6Storage Management with DB2 for OS/390
• DB2 data sharing performance improvement for open/closeof data sets
(especially beneficialduring DB2 star t-up)with Enhanced Catalog Sharing
(ECS); ECS reducesthe path length and supports the ICF shared catalog
on the coupling facility.
Youcan check the Appendix section G.4, “Web Sites” on page 242 for sites on
DB2 and DF SMS reporting themost current information on the supported
functions.

2.1.2 Managing DB2 Data Sets with DFSMS

The DB2 adminstratorcan use DFSMS to achieve all the objectives for dataset
placement and design.D FSMS hast he necessary flexibilityto suppor t everything
the DB2 administrator maywant. There is no reasonwhatsoever for not taking
advantage of DFSMS for DB2dat a sets.
Toachieve a successful implementation, an agreement between the storage
administratorand the DB2 administrator is required so that they can together
establishan environment that satisfies both their objectives.

2.1.3 Examples for Managing DB2 Data Sets with DFSMS

Examples are shown todes cribe one possibleway to manage DB2 data sets with
DFSMS. These examplesare not supposedto be a recommendation. The
examples are shownt o givean idea on t he possibilitiesthat DFSMS offers for
DB2. Each example is just one out of many choices of how a medium to complex
installation mayapproach the implementation of DB2 data sets with DFSMS.
Many installations mayfind a simpler implementation more adequate, while
others may wantto have a more specific management than the one s hown.
2.2 DB2 and Storage Servers
DB2 has some special requirements in the wayits storage objects are defined
and utilized. Disk technologyhas evolved introducing RAID architecture, large
cache, virtual architecture.DBAs and storage administrators need to agree on
common actions in orderto take advantage of the available enhancements.

2.2.1 Data Placement

With smaller disk devices,without cache, data locality was important for
performance,to reduce seek and rotation times. The new disk architectures, with
concepts likelog structured files and with cache in the gigabyte sizes, havea
noticeable impact on database physicaldesign considerations. Conventional
database design rulesbased on data set placement are becoming less important
and can be ignored in most cases.

2.2.2 Large Cache

Most storage servers with large cache (greater than 1 GB) ignore the bypass
cache or inhibit cache load requests from the application. They always use the
cache; however,they continue to take into account the specifications of usage
from the applications by just scaling down or up the track retention into the cache
for reuse.