660 likes | 805 Views
Adaptive Query Processing. Adapted from a Tutorial given at SIGMOD 2006 by: Amol Deshpande, University of Maryland Joseph M. Hellerstein, University of California, Berkeley Vijayshankar Raman, IBM Almaden Research Center. Data Independence Redux. d app. d env. <<. dt. dt.
E N D
Adaptive Query Processing Adapted from a Tutorial given at SIGMOD 2006 by: Amol Deshpande, University of Maryland Joseph M. Hellerstein, University of California, Berkeley Vijayshankar Raman, IBM Almaden Research Center
Data Independence Redux dapp denv << dt dt • The taproot of modern database technology • Separation of specification (“what”) from implementation (“how”) • Refamiliarizing ourselves: Why do we care about data independence?
D. I. Adaptivity • Query Optimization: the key to data independence • bridges specification and implementation • isolates static applications from dynamic environments • How does a DBMS account for dynamics in the environment? • This tutorial is on a 30-year-old topic • With a 21st-Century renaissance ADAPTIVITY
Why the Renaissance? ? denv dt • Breakdown of traditional query optimization • Queries over many tables • Unreliability of traditional cost estimation • Success & maturity make problems more apparent, critical • c.f. Oracle v6! • Query processing in new environments • E.g. data integration, web services, streams, P2P, sensornets, hosting, etc. • Unknown and dynamic characteristics for data and runtime • Increasingly aggressive sharing of resources and computation • Interactivity in query processing • Note two separate themes? • Unknowns: even static properties often unknown in new environments • and often unknowable a priori • Dynamics: can be very high -- motivates intra-query adaptivity
20th Century Summary • System R’s optimization scheme deemed the winner for 25 years • Nearly all 20thC research varied System R’s individual steps • More efficient measurement (e.g. sampling) • More efficient/effective models (samples, histograms, sketches) • Expanded plan spaces (new operators, bushy trees, richer queries and data models, materialized views, parallelism, remote data sources, etc) • Alternative planning strategies (heuristic and enumerative) • Speaks to the strength of the scheme • independent innovation on multiple fronts • as compared with tight coupling of INGRES • But… minimal focus on the interrelationship of the steps • Which, as we saw from Ingres, also affects the plan space
21st Century Adaptive Query Processing • (well, starts in late 1990’s) • Revisit basic architecture of System R • In effect, change the basic adaptivity loop! • As you examine schemes, keep an eye on: • Rate of change in the environment that is targeted • How radical the scheme is wrt the System R scheme • ease of evolutionary change • Increase in plan space: are there new, important opportunities? • even if environment is ostensibly static! • New overheads introduced • How amenable the scheme is to independent innovation at each step • Measure/Analyze/Plan/Actuate
Tangentially Related Work • An incomplete list!!! • Competitive Optimization [Antoshenkov93] • Choose multiple plans, run in parallel for a time, let the most promising finish • 1x feedback: execution doesn’t affect planning after the competition • Parametric Query Optimization [INSS92, CG94, etc.] • Given partial stats in advance. Do some planning and prune the space. At runtime, given the rest of statistics, quickly finish planning. • Changes interaction of Measure/Model and Planning • No feedback whatsoever, so nothing to adapt to! • “Self-Tuning”/“Autonomic” Optimizers [CR94, CN97, BC02, etc.] • Measure query execution (e.g. cardinalities, etc.) • Enhances measurement, on its own doesn’t change the loop • Consider building non-existent physical access paths (e.g. indexes, partitions) • In some senses a separate loop – adaptive database design • Longer timescales
Tangentially Related Work II • Robust Query Optimization [CHG02, MRS+04, BC05, etc.] • Goals: • Pick plans that remain predictable across wide ranges of scenarios • Pick least expected cost plan • Changes cost function for planning, not necessarily the loop. • If such functions are used in adaptive schemes, less fluctuation [MRS+04] • Hence fewer adaptations, less adaptation overhead • Adaptive query operators [NKT88, KNT89, PCL93a, PCL93b] • E.g. memory-adaptive sort and hash-join • Doesn’t address whole-query optimization problems • However, if used with AQP, can result in complex feedback loops • Especially if their actions affect each other’s models!
Extended Topics in Adaptive QP • An incomplete list!! • Parallelism & Distribution • River [A-D03] • FLuX [SHCF03, SHB04] • Distributed eddies [TD03] • Data Streams • Adaptive load shedding • Shared query processing
Adaptive Selection Ordering Title slide
Selection Ordering • Complex predicates on relations common • Eg., on an employee relation: ((salary > 120000) AND (status = 2)) OR ((salary between 90000 and 120000) AND (age < 30) AND (status = 1)) OR … • Selection ordering problem Decide the order in which to evaluate the individual predicates against the tuples • We focus on evaluating conjunctive predicates (containing only AND’s) Example Query select * from R where R.a = 10 and R.b < 20 and R.c like ‘%name%’;
Why Study Selection Ordering • Many join queries reduce to this problem • Queries posed against a star schema • Queries where only pipelined left-deep plans are considered • Queries involving web indexes • Increasing interest in recent years • Web indexes [CDY’95, EHJKMW’96, GW’00] • Web services [SMWM’06] • Data streams [AH’00, BMMNW’04] • Sensor Networks [DGMH’05] • Similar to many problems in other domains • Sequential testing (e.g. for fault detection) [SF’01, K’01] • Learning with attribute costs [KKM’05]
Execution Strategies result R • For each tuple r Є R • Apply predicate R.a = 10 first; • If tuple satisfies the selection, apply R.b < 20; • If both satisfied, apply R.c like ‘%name%’; R.c like … R.a = 10 R.b < 20 Pipelined execution (tuple-at-a-time) Static optimization ? 1. Using the KBZ algorithm Order by c/(1-p) Assumes predicate independence 2. A greedy algorithm Known to be 4-approximate
Adaptive Greedy [BMMNW’04] R1 R2 R3 R.c like … R.c like … R.a = 10 R.a = 10 R.b < 20 R.b < 20 R result Costs 1 unit 1 unit 1 unit Initial estimated selectivities 0.05 0.1 0.2 • Context: Pipelined query plans over streaming data • Example: Three independent predicates Optimal execution plan orders by selectivities (because costs are identical)
Adaptive Greedy [BMMNW’04] R1 R2 R3 R.c like … R.a = 10 R.b < 20 R result Profile • Monitor the selectivities • Switch order if the predicates not ordered by selectivities Randomly sample R.a = 10 estimate selectivities of the predicates over the tuples of the profile R.b < 20 Rsample R.c like … Reoptimizer IF the current plan not optimal w.r.t. these new selectivities THEN reoptimize using the Profile
Adaptive Greedy [BMMNW’04] R1 Randomly sample R2 R3 R.c like … R.a = 10 R.b < 20 R result R.a = 10 monitor selectivities sel(R.a = 10), sel(R.b < 20), sel(R.c …) R.b < 20 Rsample (Profile) R.c like … monitor conditional selectivities sel(R.b < 20 | R.a = 10) sel(R.c like … | R.a = 10) sel(R.c like … | R.a = 10 and R.b < 20) • Correlated Selections • Must monitor conditional selectivities Reoptimizer Uses conditional selectivities to detect violations Uses the profile to reoptimize O(n2) selectivities need to be monitored
Adaptive Greedy [BMMNW’04] • Advantages: • Can adapt very rapidly • Theoretical guarantees on performance • Not known for any other AQP protocols • Disadvantages: • Limited applicability • Only applies to selection ordering and specific types of join queries • Possibly high runtime overheads • Several heuristics described in the paper
Eddies [AH’00] A traditional pipelined query plan R1 R2 R3 R.c like … R.a = 10 R.b < 20 R result R.a = 10 Eddy R.b < 20 R result R.c like … • Treat query processing as routing of tuples through operators Pipelined query execution using an eddy • An eddy operator • Intercepts tuples from sources and output tuples from operators • Executes query by routing source • tuples through operators
Eddies [AH’00] R.a = 10 Eddy R.b < 20 R result R.c like … An R Tuple: r1 r1 r1
Eddies [AH’00] Operator 1 ready bit i : 1 operator i can be applied 0 operator i can’t be applied R.a = 10 Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 An R Tuple: r1 r1
Eddies [AH’00] Operator 1 done bit i : 1 operator i has been applied 0 operator i hasn’t been applied R.a = 10 Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 An R Tuple: r1 r1
Eddies [AH’00] Operator 1 R.a = 10 Used to decide validity and need of applying operators Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 An R Tuple: r1 r1
Eddies [AH’00] Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 An R Tuple: r1 not satisfied For a query with only selections, ready = complement(done) r1 r1 r1 satisfied r1 eddy looks at the next tuple
Eddies [AH’00] Operator 1 satisfied R.a = 10 Operator 2 Eddy R.b < 20 satisfied R result R.c like … Operator 3 satisfied An R Tuple: r2 r2
Eddies [AH’00] Operator 1 satisfied R.a = 10 Operator 2 Eddy R.b < 20 satisfied R result R.c like … Operator 3 satisfied An R Tuple: r2 if done = 111, send to output r2 r2
Eddies [AH’00] Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 • Adapting order is easy • Just change the operators to which tuples are sent • Can be done on a per-tuple basis • Can be done in the middle of tuple’s “pipeline” • How are the routing decisions made ? • Using a routing policy
Routing Policy 1: Non-adaptive table lookups very efficient Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 • Simulating a single static order • E.g. operator 1, then operator 2, then operator 3 Routing policy: if done = 000 route to 1 100 route to 2 110 route to 3
Overhead of Routing • PostgreSQL implementation of eddies using bitset lookups [Telegraph Project] • Queries with 3 selections, of varying cost • Routing policy uses a single static order, i.e., no adaptation
Routing Policy 2: Deterministic Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 • Monitor costs and selectivities continuously • Reoptimize periodically using KBZ Can use the A-Greedy policy for correlated predicates Statistics Maintained: Costs of operators Selectivities of operators Routing policy: Use a single order for a batch of tuples Periodically apply KBZ
Overhead of Routing and Reoptimization • Adaptation using batching • Reoptimized every X tuples using monitored selectivities • Identical selectivities throughout experiment measures only the overhead
Routing Policy 3: Lottery Scheduling Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 R result R.c like … Operator 3 • Originally suggested routing policy [AH’00] • Applicable when each operator runs in a separate “thread” • Can also be done single-threaded, via an event-driven query executor • Uses two easily obtainable pieces of information for making routing decisions: • Busy/idle status of operators • Tickets per operator
Routing Policy 3: Lottery Scheduling Rule: IF operator busy, THEN do not route more tuples to it BUSY Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 IDLE R result R.c like … Operator 3 IDLE • Routing decisions based on busy/idle status of operators Rationale: Every thread gets equal time SO IF an operator is busy, THEN its cost is perhaps very high
Routing Policy 3: Lottery Scheduling Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 result R.c like … Operator 3 • Routing decisions based on tickets Rules: 1. Route a new tuple randomly weighted according to the number of tickets tickets(O1) = 10 tickets(O2) = 70 tickets(O3) = 20 Will be routed to: O1 w.p. 0.1 O2 w.p. 0.7 O3 w.p. 0.2 r
Routing Policy 3: Lottery Scheduling Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 result R.c like … Operator 3 • Routing decisions based on tickets Rules: 1. Route a new tuple randomly weighted according to the number of tickets tickets(O1) = 10 tickets(O2) = 70 tickets(O3) = 20 r
Routing Policy 3: Lottery Scheduling Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 result R.c like … Operator 3 • Routing decisions based on tickets Rules: 1. Route a new tuple randomly weighted according to the number of tickets 2. route a tuple to an operator Oi tickets(Oi) ++; tickets(O1) = 11 tickets(O2) = 70 tickets(O3) = 20
Routing Policy 3: Lottery Scheduling Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 result R.c like … Operator 3 • Routing decisions based on tickets Rules: 1. Route a new tuple randomly weighted according to the number of tickets 2. route a tuple to an operator Oi tickets(Oi) ++; 3. Oi returns a tuple to eddy tickets(Oi) --; tickets(O1) = 11 tickets(O2) = 70 tickets(O3) = 20 r
Routing Policy 3: Lottery Scheduling Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 Will be routed to: O2 w.p. 0.777 O3 w.p. 0.222 result R.c like … Operator 3 • Routing decisions based on tickets Rules: 1. Route a new tuple randomly weighted according to the number of tickets 2. route a tuple to an operator Oi tickets(Oi) ++; 3. Oi returns a tuple to eddy tickets(Oi) --; tickets(O1) = 10 tickets(O2) = 70 tickets(O3) = 20 r
Routing Policy 3: Lottery Scheduling Operator 1 R.a = 10 Operator 2 Eddy R.b < 20 result R.c like … Operator 3 • Routing decisions based on tickets Rules: 1. Route a new tuple randomly weighted according to the number of tickets 2. route a tuple to an operator Oi tickets(Oi) ++; 3. Oi returns a tuple to eddy tickets(Oi) --; tickets(O1) = 10 tickets(O2) = 70 tickets(O3) = 20 Rationale: Tickets(Oi) roughly corresponds to (1 - selectivity(Oi)) So more tuples are routed to the more selective operators
Routing Policy 3: Lottery Scheduling • Effect of the combined lottery scheduling policy: • Low cost operators get more tuples • Highly selective operators get more tuples • Some tuples are randomly, knowingly routed according to sub-optimal orders • To explore • Necessary to detect selectivity changes over time
Routing Policy 4: Content-based Routing Operator 1 Eddy result Operator 2 • Routing decisions made based on the values of the attributes [BBDW’05] • Also called “conditional planning” in a static setting [DGHM’05] • Less useful unless the predicates are expensive • At the least, more expensive than r.d > 100 Expensive predicates Example Eddy notices: R.d > 100 sel(op1) > sel(op2) & R.d < 100 sel(op1) < sel(op2) Routing decisions for new tuple “r”: IF (r.d > 100): Route to op1 first w.h.p ELSE Route to op2 first w.h.p
Eddies: Post-Mortem • Cost of adaptivity • Routing overheads • Minimal with careful engineering • E.g. using bitset-indexed routing arrays • “Batching” helps tremendously • Statistics Maintenance • Executing the routing policy logic
Discussion • Benefits for AQP techniques come from two places • Increased explored plan space • Can use different plans for different parts of data • Adaptation • Can change the plan according to changing data characteristics • Selection ordering is STATELESS • No inherent “cost of switching plans” • Can switch the plan without worrying about operator states • Key to the simplicity of the techniques
Discussion • Adaptation is not free • Costs of monitoring and maintaining statistics can be very high • A selection operation may take only 1 instruction to execute • Comparable to updating a count • “Sufficient statistics” • May need to maintain only a small set of statistics to detect violations • E.g. The O(n2) matrix in Adaptive-Greedy [BBMNW’04]
Adaptive Join Processing Title slide
Outline • 20th Century Adaptivity: Intuition from the Classical Systems • Adaptive Selection Ordering • Adaptive Join Processing • Additional complexities beyond selection ordering • Four plan spaces • Simplest: pipelines of Nested Loop Joins • Traditional: Trees of Binary Operators (TOBO) • Multi-TOBO: horizontal partitioning • Dataflows of unary operators • Handling asynchrony • Research Roundup
Select-Project-Join Processing SMJ NLJ T S R • Query: select count(*) from R, S, T where R.a=S.aandS.b=T.b and S.c like ‘%name%’ and T.d = 10 • An execution plan • Cost metric: CPU + I/O • Plan Space: • Traditionally, tree of binary join operators (TOBO): • access methods • Join algorithms • Join order • Adaptive systems: • Some use the same plan space, but switch between plans during execution • Others use much larger plan spaces • different adaptation techniques adapt within different plan spaces
Approach NLJ NLJ NLJ B C A • Pipelined nested-loops plans • Static Rank Ordering • Dynamic Rank Ordering – Eddies, Competition • Trees of Binary Join Operators (TOBO) • Static: System R • Dynamic: Switching plans during execution Multiple Trees of Binary Join Operators • Convergent Query Processing • Eddies with Binary Join Operators • STAIRs: Moving state across joins • Dataflows of Unary Operators • N-way joinsswitching join algorithms during execution • Asynchrony in Query Processing
Pipelined Nested Loops Join R NLJ NLJ NLJ B C A • Simplest method of joining tables • Pick a driver table (R). Call the rest driven tables • Pick access methods (AMs) on the driven tables • Order the driven tables • Flow R tuples through the driven tables For each r R do:look for matches for r in A;for each match a do: look for matches for <r,a> in B; …
Adapting a Pipelined Nested Loops Join NLJ NLJ NLJ B C A Almost identical to selection ordering • Simplest method of joining tables • Pick a driver table (R). Call the rest driven tables • Pick access methods (AMs) on the driven tables • Order the driven tables • Flow R tuples through the driven tables For each r R do:look for matches for r in A;for each match a do: look for matches for <r,a> in B; … R Keep this fixed for now
Tree Of Binary Join Operators (TOBO) recap • Standard plan space considered by most DBMSs today • Pick access methods, join order, join algorithms • search in this plan space done by dynamic programming NLJ MJ HJ C D NLJ B A R