Blog de Ronald Vasseur

Un article de Webman    Pas de commentaires

Voici une petite news concernant SQL Server 2008, nom de code Katmai. Voici la liste, je pense quasi exhaustive, des nouveautés apportées par SQL Server 2008, dont le lancement est prévu en février 2008 en meme temps que Visual Studio 2008 et Windows Server 2008. Cette liste provient du blog de Chad Boyd, un ingénieur SQL Server de chez Microsoft, donc je pense que nous pouvons la considérer comme fiable. En voici la copie :

Security/Auditing
– Transparent Data Encryption (encryption while data is ‘still’ on disk, transparent to applications)
– External Key Management (Consolidation of key management, integration with external products)
– Data Auditing (1st-class ‘AUDIT’ objects; DDL support; audit objects, principals, data, etc.; support for multiple logging targets)

Availability/Reliability
– Pluggable CPU support
– Enhanced Database Mirroring (compression of mirror streams, enhanced performance, automatic page-level repair for principal/mirror)

Performance
– Data compression (easy to enable/disable online, more efficient data storage (this is NOT traditional data compression))
– Backup stream compression (server level control or backup statement control, all backup types)
– Performance data collection (single, common framework for data collection, reporting, and storage/warehousing)
– Improved Plan Guide support (plan freezing, pull plans directly from plan cache, SSMS integration, etc.)
– Resource Governor (create pools and groups to govern, define classifications based on built-in functions, segment resource utilization amoung groups)

Management
– Policy-based management framework (manage via policies vs. scripts, enterprise-wide support, automated monitoring/enforcement, etc.)
– Integrate with Microsoft System Center
– Extended Events (high perf lightweight tracing infrastructure, NOT sql trace, integrated with ETW, unprecidented insight into goings-on)

Development Enhancements
– Improved datetime datatypes (100th nanosecond precision (7 digits past second), time-zone datetime offset, date only, time only)
– HierarchyID datatype (hierarchical-aware data type, ORDPath values, built-in functions, methods, etc.)
– Entity Data Model support (develop ‘business entities’ vs. tables, model complex relationships, retrieve entities vs. rows/columns)
– LINQ
– Sql Server Change Tracking (Change Data Capture, get ‘diff’ data changes WITHOUT a comparible value (i.e. datetime, timestamp, etc.))
– Table Valued Parameters
– MERGE statement (‘upsert’ data, also includes deletion functionality)
– Large UDT’s (no more 8000 byte limit on CLR-based UDTs, no more 8000 byte limit for UDA’s)
– Spatial data (GEOMETRY and GEOGRAPHY data types, built-in spatial function support, spatial indexes)
– XML enhancements (support for lax validation, office 12 support, xs:dateTime support, lists/union types, LET FLOWR support, etc.)
– Inline initialization and compound assignment

Service Broker
– New UI and Tools for working with (add/drop/edit functionality within SSMS, Diag tools, )
– Conversation Priority (set message ordering, send/receive impact, 1-10 levels)

Data Storage
– Data compression (see above)
– FILESTREAM attribute (get the ‘best of both’ functionality from BLOBs in the DB vs. BLOBs on filesystem, no more « to blob or not to blob »)
– Integrated Full Text Search (FTS fully integrated into DB engine, no external storage, no external service, more efficient and reliable costing)
– Sparse columns (more efficient storage for ‘wide’ tables with many columns that repeat and don’t contain data)
– New index types (spatial indexes, hierarchical indexes, FILTERED indexes (indexes on filtered values within columns), etc.)

Data Warehousing/ETL
– Partitioned Table Parallelism (no more thread limit per partition)
– Star Join support (no special syntax, optimizer based, full backward syntax support)
– Data compression (see above)
– Resource Governor (see above)
– Persistent Lookups in SSIS (no more re-querying for lookup operators, cache lookups in multiple ways, persist lookups to disk)
– Improved thread scheduling in SSIS (shared thread pool, pipeline parallelism)
– Change Data Capture (see above)
– MERGE statement (see above, great uses with slowly changing dimensions)
– Scale-out analysis services (read-only storage supports multiple AS servers)
– Subspace computations
– New Tools for Cube design
– Best Practice Design Alerting
– Backup cubes with better scalability
– Data-mining add-ins for Excell

Reporting
– IIS Agnostic Reporting Services Deployment (no IIS required to run RS any longer)
– Rich-text support
– Enhanced visualiztion (graphing)
– New Word rendering (render reports to Microsoft Word)

Deprecation
– Many ‘old’ features ARE REMOVED/GONE (those that have been deprecated for some time – 60/65/70 compat modes, nolog / truncateonly syntax, etc.)

En tant que développeur on appreciera particulièrment la suppression de la limitation à 8000 octets des User Defined Types, l’encryptage de données qui est simplifié et peut être totalement transparent pour les utilisateurs/développeurs, et le support de LINQ (qui sera une l’une, si ce n’est la plus importante nouveauté du Framework .Net 3.5). Il me tarde le mois de février pour avoir accès à toutes ces « petites choses » qui rendent la vie de développeur plus interessante !! Pour information je vous rappelle qu’Orcas (Visual Studio 2008) sera disponible vers la fin de l’année 2007, bien que son lancement officiel interviendra lui, en février 2008.