Class ThrottleBin
- java.lang.Object
-
- org.apache.manifoldcf.connectorcommon.throttler.ThrottleBin
-
public class ThrottleBin extends java.lang.Object
Throttles for a bin. An instance of this class keeps track of the information needed to bandwidth throttle access to a url belonging to a specific bin. In order to calculate the effective "burst" fetches per second and bytes per second, we need to have some idea what the window is. For example, a long hiatus from fetching could cause overuse of the server when fetching resumes, if the window length is too long. One solution to this problem would be to keep a list of the individual fetches as records. Then, we could "expire" a fetch by discarding the old record. However, this is quite memory consumptive for all but the smallest intervals. Another, better, solution is to hook into the start and end of individual fetches. These will, presumably, occur at the fastest possible rate without long pauses spent doing something else. The only complication is that fetches may well overlap, so we need to "reference count" the fetches to know when to reset the counters. For "fetches per second", we can simply make sure we "schedule" the next fetch at an appropriate time, rather than keep records around. The overall rate may therefore be somewhat less than the specified rate, but that's perfectly acceptable. Some notes on the algorithms used to limit server bandwidth impact ================================================================== In a single connection case, the algorithm we'd want to use works like this. On the first chunk of a series, the total length of time and the number of bytes are recorded. Then, prior to each subsequent chunk, a calculation is done which attempts to hit the bandwidth target by the end of the chunk read, using the rate of the first chunk access as a way of estimating how long it will take to fetch those next n bytes. For a multi-connection case, which this is, it's harder to either come up with a good maximum bandwidth estimate, and harder still to "hit the target", because simultaneous fetches will intrude. The strategy is therefore: 1) The first chunk of any series should proceed without interference from other connections to the same server. The goal here is to get a decent quality estimate without any possibility of overwhelming the server. 2) The bandwidth of the first chunk is treated as the "maximum bandwidth per connection". That is, if other connections are going on, we can presume that each connection will use at most the bandwidth that the first fetch took. Thus, by generating end-time estimates based on this number, we are actually being conservative and using less server bandwidth. 3) For chunks that have started but not finished, we keep track of their size and estimated elapsed time in order to schedule when new chunks from other connections can start. NOTE WELL: This is entirely local in operation
-
-
Nested Class Summary
Nested Classes Modifier and Type Class Description protected static class
ThrottleBin.SumClass
-
Field Summary
Fields Modifier and Type Field Description protected java.lang.String
binName
This is the bin name which this throttle belongs to.protected boolean
estimateInProgress
Flag indicating whether rate estimation is in progress yetprotected boolean
estimateValid
Flag indicating whether a rate estimate is neededprotected boolean
isAlive
This signals whether the bin is alive or not.protected double
localMinimum
The local minimum milliseconds per byteprotected double
minimumMillisecondsPerByte
The minimum milliseconds per byteprotected double
rateEstimate
The inverse rate estimate of the first fetch, in ms/byteprotected int
refCount
This is the reference count for this bin (which records active references)protected long
seriesStartTime
The start time of this seriesprotected java.lang.String
serviceName
The (anonymous) service nameprotected java.lang.String
serviceTypeName
Service type nameprotected static java.lang.String
serviceTypePrefix
The service type prefix for throttle binsprotected java.lang.String
targetCalcLockName
The target calculation lock nameprotected static java.lang.String
targetCalcLockPrefix
The target calculation lock prefixprotected long
totalBytesRead
Total actual bytes read in this series; this includes fetches in progress
-
Constructor Summary
Constructors Constructor Description ThrottleBin(IThreadContext threadContext, java.lang.String throttlingGroupName, java.lang.String binName)
Constructor.
-
Method Summary
All Methods Static Methods Instance Methods Concrete Methods Modifier and Type Method Description void
abortFetch()
Abort the fetch.void
abortRead()
Abort a read in progress.void
beginFetch()
Note the start of a fetch operation for a bin.boolean
beginRead(int byteCount, IBreakCheck breakCheck)
Note the start of an individual byte read of a specified size.protected static java.lang.String
buildServiceTypeName(java.lang.String throttlingGroupName, java.lang.String binName)
protected static java.lang.String
buildTargetCalcLockName(java.lang.String throttlingGroupName, java.lang.String binName)
boolean
endFetch()
Note the end of a fetch operation.void
endRead(int originalCount, int actualCount)
Note the end of an individual read from the server.java.lang.String
getBinName()
Get the bin name.protected static byte[]
pack(double targetDouble)
void
poll(IThreadContext threadContext)
Poll this binvoid
shutDown(IThreadContext threadContext)
Shut down this bin.protected static double
unpackTarget(byte[] data)
void
updateMinimumMillisecondsPerByte(double min)
Update minimumMillisecondsPerBytePerServer
-
-
-
Field Detail
-
isAlive
protected boolean isAlive
This signals whether the bin is alive or not.
-
binName
protected final java.lang.String binName
This is the bin name which this throttle belongs to.
-
serviceTypeName
protected final java.lang.String serviceTypeName
Service type name
-
serviceName
protected final java.lang.String serviceName
The (anonymous) service name
-
targetCalcLockName
protected final java.lang.String targetCalcLockName
The target calculation lock name
-
minimumMillisecondsPerByte
protected double minimumMillisecondsPerByte
The minimum milliseconds per byte
-
localMinimum
protected double localMinimum
The local minimum milliseconds per byte
-
refCount
protected volatile int refCount
This is the reference count for this bin (which records active references)
-
rateEstimate
protected double rateEstimate
The inverse rate estimate of the first fetch, in ms/byte
-
estimateValid
protected volatile boolean estimateValid
Flag indicating whether a rate estimate is needed
-
estimateInProgress
protected volatile boolean estimateInProgress
Flag indicating whether rate estimation is in progress yet
-
seriesStartTime
protected long seriesStartTime
The start time of this series
-
totalBytesRead
protected long totalBytesRead
Total actual bytes read in this series; this includes fetches in progress
-
serviceTypePrefix
protected static final java.lang.String serviceTypePrefix
The service type prefix for throttle bins- See Also:
- Constant Field Values
-
targetCalcLockPrefix
protected static final java.lang.String targetCalcLockPrefix
The target calculation lock prefix- See Also:
- Constant Field Values
-
-
Constructor Detail
-
ThrottleBin
public ThrottleBin(IThreadContext threadContext, java.lang.String throttlingGroupName, java.lang.String binName) throws ManifoldCFException
Constructor.- Throws:
ManifoldCFException
-
-
Method Detail
-
buildServiceTypeName
protected static java.lang.String buildServiceTypeName(java.lang.String throttlingGroupName, java.lang.String binName)
-
buildTargetCalcLockName
protected static java.lang.String buildTargetCalcLockName(java.lang.String throttlingGroupName, java.lang.String binName)
-
getBinName
public java.lang.String getBinName()
Get the bin name.
-
updateMinimumMillisecondsPerByte
public void updateMinimumMillisecondsPerByte(double min)
Update minimumMillisecondsPerBytePerServer
-
beginFetch
public void beginFetch()
Note the start of a fetch operation for a bin. Call this method just before the actual stream access begins. May wait until schedule allows.
-
abortFetch
public void abortFetch()
Abort the fetch.
-
beginRead
public boolean beginRead(int byteCount, IBreakCheck breakCheck) throws java.lang.InterruptedException, BreakException
Note the start of an individual byte read of a specified size. Call this method just before the read request takes place. Performs the necessary delay prior to reading specified number of bytes from the server.- Returns:
- false if the wait was interrupted due to the bin being shut down.
- Throws:
java.lang.InterruptedException
BreakException
-
abortRead
public void abortRead()
Abort a read in progress.
-
endRead
public void endRead(int originalCount, int actualCount)
Note the end of an individual read from the server. Call this just after an individual read completes. Pass the actual number of bytes read to the method.
-
endFetch
public boolean endFetch()
Note the end of a fetch operation. Call this method just after the fetch completes.
-
poll
public void poll(IThreadContext threadContext) throws ManifoldCFException
Poll this bin- Throws:
ManifoldCFException
-
shutDown
public void shutDown(IThreadContext threadContext) throws ManifoldCFException
Shut down this bin.- Throws:
ManifoldCFException
-
unpackTarget
protected static double unpackTarget(byte[] data)
-
pack
protected static byte[] pack(double targetDouble)
-
-