Transcript SQL Service Broker

SQL Service Broker
Gerald Hinson
Development Lead – SSB team
HPTS 2005
SQL Service Broker
Slide 1
THESIS: Queues are Databases [1]
Jim Gray - December 21, 1995
Abstract:
Message-oriented-middleware (MOM) has become an small industry.
MOM offers queued transaction processing as an advance over pure
client-server transaction processing. This note makes four points:
– Queued transaction processing is less general than direct transaction
processing.
– Queues are interesting databases with interesting concurrency
control.
– Queue systems need DBMS functionality.
– Queue managers are simple TP-monitors managing server pools driven
by queues.
[1] Acknowledgments:
These ideas derive from discussions with Andrea Borr (Oracle), Richard Carr
(Tandem), Dieter Gawlick (Oracle), Pat Helland (Microsoft), Franco Putzolu
(Oracle), Andreas Reuter (U. Stuttgart) and Bill Highliman (NetWeave).
SQL Service Broker
Slide 2
THESIS: Queues are Databases [1]
Jim Gray - December 21, 1995
 Just for fun, here’s my favorite quote from the paper:
“This position paper was intended to generate controversy at
the High Performance Transaction Processing Workshop
(HPTS). Amazingly, everyone either agreed or was so
disgusted that they left the room. In the end, there was no
heated discussion. I was astonished.”
SQL Service Broker
Slide 3
SQL Server 2005 - Mild-mannered database
becomes App Server
 What did we add?
– Reliable messaging, but with a new twist…
– Application activation
• Launching of either stored procs or .EXEs to consume
queued messages
– CLR hosting for VB, C# based stored procs
– Better XML support (XPath / XQuery)
 Why?
– Customer demand
• They weren’t happy with the separate MOM either…
– Dieter did it for Oracle
– I still remember IMS DB/DC (and DB/2)
SQL Service Broker
Slide 4
1 of everything vs. 2 of everything
 When messaging is merged into the database
customers get:
– 1 programming model
– 1 product to install, configure, monitor
– 1 backup story, 1 failover story
• Because ALL app data (queues, session state, etc.)
lives in a single database
– 1 data warehousing story
– Queued messages and application session state
may be queried against
– Simplified deployment of applications
• Cool replication synergies here
It’s a natural pattern.
– Customers keep building it themselves (vs. using dedicated
queuing products such as MSMQ, MQ-Series, etc.)
SQL Service Broker
Slide 5
Rationale for a new communication primitive
 Programming Model issues :
– This is a typical IT scenario and its typical
solution.
Billing Program
Send PO Request
Loop until EOD msg
Recv Line Item
Insert into…
End-Loop
Finalize Bill
PO Program
Recv PO request
Select * from PO_tbl
Loop until done
Fetch Line Item
Send Line Item
End-Loop
Send EOD msg
– This is the obvious, simple solution.. But, it
doesn’t work…
SQL Service Broker
Slide 6
Rationale for a new communication primitive
(cont.)
Billing Program
(Instance 1)
Billing Program
(Instance 2)
Billing
Database
Billing Program
(Instance 3)
This one runs first
and finalizes the
bill early
Billing Queue
Line Item 1
Line Item 2
Line Item 3
EOD msg
End-of-Q
Incoming PO
data from
PO Service
 Adding multiple instances to drain the queue faster breaks the
app!
– Exceptions, deadlocks and contention are all common outcomes.
App programmers get burned by this a lot using existing
message queuing solutions.
– The typical workaround is to force all data into a single
message… But, that leads issues with message size, latency,
etc.
SQL Service Broker
Slide 7
Rationale for a new communication primitive
(cont.)
 Programming Model issues :
– Weak correlation of related messages
– No support for relating messages and application state
Hotel
Client
Travel
Travel
Travel
Airline
Car Rental
DB
Single Business Transaction
– In the example above, each instance of the Travel
service may deadlock another trying to process the
responses from the back-end services.
SQL Service Broker
Slide 8
Rationale for a new communication primitive
(cont.)
Travel Program
(Instance 1)
Travel Program
(Instance 2)
Travel
Database
Travel Program
(Instance 3)
Each instance of
“Travel” tries to
work on the same
business tran.
Travel Queue
Hotel Reply
Airline Reply
Car Reply
End-of-Q
Incoming
replies from
back-end
services.
 Adding multiple instances to drain the queue faster
breaks the app here too!
– Composition of services becomes too hard for application
programmers… (SELECT, followed by app locks, followed
by…)
SQL Service Broker
Slide 9
Rationale for new communication primitive
(cont.)
 Programming model conclusions:
– Building composable message-based services is
harder than it should be.
• Atomic messages, as a communication
primitive, make building composable services
hard.
– Similar lessons were learned at the network
layer
• UDP has become a niche solution. TCP’s added
value makes like simpler.
SQL Service Broker
Slide 10
Dialogs - A new communication primitive
 SQL Server apps (read: Services) communicate via formal,
reliable sessions known as Dialogs
 Services communicate with other services by beginning a
dialog and sending / receiving messages on it. Examples
of new DML:
– BEGIN DIALOG [CONVERSATION] @dialogHandle
FROM SERVICE [CustomerService]
TO SERVICE ‘TravelService'
- SEND ON CONVERSATION @dialogHandle
MESSAGE TYPE [BookTravelRequest] …
- RECEIVE * from MyQueue WHERE …
or if you want to block…
WAITFOR (RECEIVE * from MyQueue), TIMEOUT = milliseconds
- RECEIVE returns spooled message for one or more dialogs
- END CONVERSATION @dialogHandle
SQL Service Broker
Slide 11
Dialogs – A new communication primitive
(cont.)
 Dialogs provide two-way messaging between two services
 Dialogs offer:
–
–
–
–
Guaranteed delivery
Full-duplex or half-duplex communication
Exactly-once, in-order delivery
Private / public key based security
 Dialogs:
– May be long-lived (years) or short-lived (seconds)
– Are light-weight
– Are sessions (similar to TCP conceptually)
Customer
Service
Travel
Service
Dialog
Database A
SQL Service Broker
Database B
Slide 12
Conversation Groups
ConvGroup1
ConvGroup1
ConvGroup1
ConvGroup1
Client
Dialog 1
Dialog1
Dialog2
Dialog3
Dialog4
Travel
Dialogs 2,3,4 share a Conversation Group with
Dialog 1 because the Travel program chose to
relate them at dialog creation time.
Hotel
Dialog 3
Airline
Car
When a transaction locks one of them (at dequeue
time) it locks them all. They are unlocked when
the transaction is committed.
The Conversation Group is owned and localized to
the Travel program. This identity does not travel
on the “wire”.
SQL Service Broker
Slide 13
Conversation Groups
 Conversation Groups:
– Formalize the relationship between dialogs
– Have a unique identity
– Are a serializable unit in SQL Server. Only one transaction may
lock a conversation group at a time.
– Dialogs are always associated with a conversation group, hence
only one transaction can lock a dialog at a time as well.
– Provide a great key (identity) to associate with application state
 If there are > 1 readers on a given queue, they will never
accidentally process related messages, dialogs or (if the
developer uses this key correctly) application state.
 Conversation Groups are meant to:
– Simplify application state management
– Enable “business transactions”
 Conversations are grouped using an optional expression on
the BEGIN DIALOG verb (RELATE TO…).
SQL Service Broker
Slide 14
Composed Services via dialogs
Travel Program
(Instance 1)
Travel Program
(Instance 2)
Travel
Database
Travel Program
(Instance 3)
Only Instance 1
can dequeue the
messages…
Travel Queue
Hotel
Airline
Car
Hotel
Airline
Car
Dialog2 ConvGroup1
Dialog3 ConvGroup1
Dialog4 ConvGroup1
Dialog5 ConvGroup2
Dialog6 ConvGroup2
Dialog7 ConvGroup2
End-of-Q
 Adding multiple instances to drain the queue faster is
now safe.
– Composition of services is now trivial.
– Application state can be stored using the Conversation
Group as a key.
SQL Service Broker
Slide 15
Related messages via dialogs
Billing Program
(Instance 1)
Billing Program
(Instance 2)
Billing
Database
Billing Program
(Instance 3)
Each transaction
is isolated from
POs already locked
by another tran…
Billing Queue
Line 1
Line 2
Line 3
EOD
Line 1
Line 2
Dialog1 ConvGroup1
Dialog1 ConvGroup1
Dialog1 ConvGroup1
Dialog1 ConvGroup1
Dialog2 ConvGroup2
Dialog2 ConvGroup2
End-of-Q
 Adding multiple instances to drain the queue faster is
now safe.
– Related messages are not dequeued by different
instances
SQL Service Broker
Slide 16
Deployment Issues with existing messaging
products
– “Plumbing” issues:
• Weak delivery guarantees:
– Most don’t guarantee EOIO “end-to-end” when there
are intermediate nodes in the topology.
Msg
Broker 2
Msg
Broker 1
Msg
Broker 1
Msg
Broker 3
Msg
Broker 4
Messages arrive
in order.
Messages arrive
out-of-order!!
Msg
Broker 2
– Topology changes like this break many existing message
queuing applications!
– This becomes problematic as an organization tries to
manage large message traffic
» 10,000 workstations  1 server doesn’t work!
SQL Service Broker
Slide 17
Delivery guarantees via dialogs (ie. Dialogs
aren’t just about programming model)
Customer
Service
Travel
Service
Dialog
Msg
Broker 3
Msg Broker 1
(Database)
Msg
Broker 4
Msg Broker 2
(Database)
 Dialogs have sessions IDs that are “end-to-end, no matter how
many hops.
 All routing, retries and ACKs are done per dialog vs. per msg
broker.
 Security is per dialog as well (Asymmetric/Symmetric key hybrid).
SQL Service Broker
Slide 18
Performance Gains – subtitled per Helland
(faster than greased …)
 Avoiding 2 Phase Commit
– Messages and data
in one store
 Hardware has come a long way
– 15,000 RPM drives on desktop
computers
– Sequential writes > 50MB /
second
– 100Mb networks < 12MB
bandwidth
 Reliable messaging in SQL can
be VERY fast
– Databases achieve perf via
sequential write speeds (ie.
write-ahead logging)
– Applications can now have
EOIO semantics AND high
message throughput
SQL Service Broker
Other
message store
SQL Server
LOB
Data
Queue
Inefficient
transaction commit
SQL Server
LOB
Data
Queue
Very efficient
transaction commit
Slide 19
Two ways to scale using Service Broker
 Scale single-broker service
– Multiple “worker-bee” exe
– Consume work off a single
broker’s queue
exe
Broker
exe
...
exe
 Scale multi-broker service
– Load-balance a service type
across multiple brokers
– Each broker equivalent
– “Scale-out” brokers
– This model is great for
multi-user, read-only,
highly-replicatable data
• Examples include
catalogs, phonebooks,
price lists, etc.
SQL Service Broker
exe
Broker
...
exe
exe
Broker
Broker
...
exe
exe
Broker
...
exe
Slide 20
SQL Express is free!
 Tier 1
– 1000s of
machines
 SQL
Express
 Tier 2
– 100s of
machines
 SQL
Express /
Workgroup
 Tier 3
– 10s of
machines
 SQL Server
Enterprise
 This is all about reducing connections and
transactions from lots to few by the time
you reach the back-end
 10,000 transactions on the front-end
cannot equate to 10,000 on the back-end!
SQL Service Broker
Slide 21
Internal Activation of Services
Database
 Activation comes in two flavors:
– Internal (we built it)
– External (we enabled you to build it)
Q
Q Rdr
 Internal Activation
– A stored procedure may be associated with a Service’s queue
using the following syntax:
• CREATE QUEUE myqueue (WITH ACTIVATION STATUS = ON,
PROCEDURE = db.schema.procedurename,
EXECUTE_AS = [valid database user], MAX_READERS = n)
– NOT a trigger
• Separate Transaction, Security, and Thread
– New parallel queue readers if needed
(and not yet MAX_READERS)
– Lots of customers are leveraging this for all sort of async
database work
SQL Service Broker
Slide 22
External Activation of Services
 External Activation
– Some customer don’t want to (or cannot) leverage stored
procedures
 SQL Server can:
– Send a reliable message to a specified “notification” queue
when a new queue reader is needed
– One “notification” queue can serve many application queues
– Message only sent when queue readers are lagging
 We provide a nice sample of how to leverage
this event to build an external activator.
– This sample will ship with SQL Server.
SQL Service Broker
Slide 23