Finding Keys In Cacheismo

In the last post I talked about how to do map-reduce like stuff with cacheimso. It occurred to me that their is no method to get all the keys. The global hashmap object now supports new method getPrefixMatchingKeys which as the name implies returns keys which match a certain prefix. Typically objects are stored with following template objectType$objectKey. Thus an object created by file quota.lua with key myKey will be stored with key quota$myKey. To do something with all the quota objects:
   
local keys = getHashMap():getPrefixMatchingKeys("quota$")
for k,v in pairs(keys) do 
   print(v)
end

If you would like the quota object itself to support say getAllKeys method then ...add this to quota.lua in scripts directory.



function getAllKeys() 

    local keys = getHashMap():getPrefixMatchingKeys("quota$")
    local result = ""
    for k,v in pairs(keys) do 
       result..v.."\n"  
    end
end

Now a get request with key quota:getAllKeys would return a new line separated list of all active quota objects in the map.

This is good enough but probably not very interesting. I am planning to support indexes as first class objects to make it faster to reach interesting objects in quick time. These indexes will automatically remove objects which are deleted or forced to expire because of LRU caching. So if you need to find all the quota objects that are close to quota limit, create an index on quota used value.

Using Cacheismo Cluster for Real Time Analytics

Cacheismo now supports invoking requests on other server.
  - getFromServer
    This function takes the server name and a key as an argument and returns the result to the script.
  - getInParallel
    This function takes map of server names to list of keys and gets values for all the keys from all the servers in parallel. Once all the results are received, the script gets a map of server names to map of keys and received values.

Here is a simple example code which gets to top accessed keys from the cacheismo cluster using a simplified, framework-less map-reduce.  The contents below should belong to file names mapreduce.lua

-- function called on every cacheismo node to get 
-- the top keys by access count
-- virtual key : mapreduce:topKeys:


function topKeys(count) 
    local keys  = accessCounter:getKeyAccessCount()
    -- sort keys by access count
    table.sort (keys)
    local result = {}
    local i      = 1
    for (i in 1..count) do 
       result[i] = keys[i]
    end
    return result
end


-- function called on one of the cacheismo nodes 
-- which distributes work to all other nodes using getInParallel
-- virtual key : mapreduce:manageGetTop:10 


function manageGetTop(count) 
     local query = {}
     for k,v in pairs(servers) do 
        query[k] = "mapreduce:topKeys:"..count
     end
     local results = getInParallel(query)
     local total = {}
     for k,v in pairs(results) do 
           for k1,v1 in pairs(v) do 
              total[k1] = v1
           end
     end
     table.sort (total)
     local topkeys = {}
     for (i in 1..count) do 
       topkeys[i] = total[i]
     end
end


I have taken some liberties with the syntax and their is no accessCounter object by default in cacheismo, but is fairly easy to create in the script.  Note that the above implementation doesn't have ant intermediate nodes, but is trivial to create by calling get for mapreduce:manageGetTop instead of calling mapreduce:topKeys.

Single cacheimso server might be serving thousands of such queries because everything is happening in non-blocking manner but we give user the illusion of synchronous stack using lua coroutines.

Links:
Cacheismo Introduction
Cacheismo Lua API
Cacheimso Sliding Window Counter Implementation in Lua
Cacheismo Quota Implementation in Lua
Cacheismo Memory Allocation
Cacheismo Source Code
Cacheismo Support

Multi-threaded vs Multi-process

Cacheismo doesn't use threads. So does redis. Memcached is multi-threaded. How to choose?
Multi-threaded applications are hard to code and understand. Non-blocking multi-threaded applications are close to nightmare. I had the honor of writing multi-threaded non-blocking http proxy @apigee. It was written in c, had JNI bindings with java and could talk to our custom interface with kernel using shared memory (to avoid kernel space to user space data copy).  Having done that, I choose not to use threads in cacheismo.

Scalability of multi-threaded applications with multiple cores is simply a function of how much work can be done in parallel. The lesser the conflicts between the threads, better is the performance. The first problem is distribution of work among the requests. If multiple threads are working  concurrently on same piece of data, it is hard for them to make progress. Memcached uses a single selector thread to bind connections to different threads. It is not optimal. It will become bottleneck when you have thousands of clients trying to connect to the server at the same time. May be all connections of one thread are sleeping whereas all connections of some other thread are killing the cpu. So even though we have some work, few threads are not bothered because it doesn't belongs to them. This is kind of not possible with memcached (it hardly uses cpu), but very much possible with cacheismo running user scripts.  Now consider the hashmap itself. One hashmap, multiple threads, unavoidable lock. Possible solution here would be striping. What about slabs. One possibility is per thread slabs or some form of thread local caching or instead of using lock at slab level, use per slab locks.

In short, to make multi-threaded code work at optimal levels, we have to looking at our objects in finer granularity than what we originally planned. We have to make our design sub optimal to make sure threads make progress, even if at the cost of under utilization of resources.
Multi-process model works the opposite way. Consider your complete code/process as an object and create multiples of them. Have some way to distribute requests to these individual objects/processes. In many cases it is possible, specifically when either clients have knowledge of servers (consistent hashing) or if some intelligent proxy can do this for us (http load balancer). Even this could be non-optimal, but at least it is simple. What else? Memory leaks are easier to manage with multi-process model. Just restart the process. Since memory is managed at process level, it can be easily reclaimed by killing the process and if we have multiple of them, the service is not disrupted for every user. In a multi-threaded model, restart will impact all users. No locks are needed in multi process code, unless we use shared memory. The process boundary gives exclusive access to memory which doesn't need further management.

I don't see multi-process code in any ways inferior to multi-threaded code. It has some benefits like simple to code, simple to debug, simple to maintain. Other beauty is if instead of running 1 process with multiple threads you are running multiple processes with single thread, crash in the first case will cause full system failure, whereas in second case you still have (n-1)/n percent of the system up.  It is easy to upgrade things one at a time because we have more than 1 thing. With single server, no option but to shut it down, upgrade and start.

The only problem I see with multi-process code is duplication of basic data structure and configuration.  For example running each requests in a separate java process in tomcat or jboss will be a really bad idea. There is simply too much shared state - db connection pool, byte code, runtime native code, session information, etc.   The other problem is a bit superficial one; instead of managing one process, we have to manage multiple of them (use scripts?)
If your solution is either stateless or it can be partitioned such that there is no interdependence, multiprocess will make more sense. It is a difficult choice, but just don't use threads if you can survive without them, specially with non-blocking code. Try  cacheismo and see for  yourself.

Cacheismo Cluster Update 2

Today I could set the value on server 1 and get it back from server 2.  This script of server 2 was changed a bit.


function handleGET(command, key) 
    local cacheItem = getHashMap():get(key)
    if (cacheItem ~= nil) then
   command:writeCacheItem(cacheItem)
   cacheItem:delete()
else 
   local result = command:getFromExternalServer("127.0.0.1:11212", key)
   if (result ~= nil) then
       writeStringAsValue(command, key, result)
   end 
    end
end

This is where I feel lua beats javascript because of its excellent coroutine support.  I could easily change the above code to a while loop and go through all servers in the cluster looking for a value and it would not change the complexity of the code a bit, but try doing it with javascript callbacks and it would look like a mess. 

I am thinking now that the original idea of using dataKey and virtualKey is a bit lame. What I would do now is that I will make server an explicit argument, just like I used above. For consistent hashing we will have an object exposed in lua which can be used to find the correct server via consistent hashing, but it would be optional to use. This gives users the flexibility to use some cacheismo instances as pure proxies and others as caches. Some of the servers can be part of one consistent hashing group and others can be part of another. For some reason if users want to use one of the servers as a naming server to map keys to server, they can use it that way. Essentially users have full flexibility with respect to how to map keys to servers  - consistent hashing, multiple servers hidden behind a proxy(cacheismo), naming server to map keys to servers or any combination of these.  

Other possibilities include replication and quorum. These can be accomplished with the current API itself, but I will add parallel get support to make it as fast as possible. 

Cacheismo Lua API

The scripting capabilities of cacheismo are very powerful. The code for handing commands like get/set/add etc is completely written in lua using the lua interface. Here is the code for handling SET command:


local function handleSET(command) 
  local hashMap   = getHashMap()   
  local cacheItem = hashMap:get(command:getKey())
  if (cacheItem ~= nil) then 
      hashMap:delete(command:getKey())
      cacheItem:delete()
      cacheItem = nil
  end 
  cacheItem = command:newCacheItem()
  if (cacheItem ~= nil) then 
      hashMap:put(cacheItem)
      command:writeString("STORED\r\n")
  else 
     command:writeString("SERVER_ERROR Not Enough Memory\r\n")
  end
  return 0
end



This is the list of the lua methods which can be called by new scripts

getHashMap()
- Global method, return instance of the global hashmap object.

setLogLevel(level)
- Sets the logging level. Four levels are defined DEBUG=0, INFO=2,
  WARN=2 and ERR=3. The function takes numeric argument.

Table 
The standard table object in lua is extended to support the following methods

marshal(table) 
- returns a string object which represents the serialized table.

unmarshal(serializedTable)
- returns a fully constructed table from the given serialized table string

clone(table) 
- returns a deep copy of the given table


HashMap
The global hashmap object retrieved from getHashMap() supports following methods

get(key) 
- returns a cacheItem object for the given key if found, else nil

put(cacheItem) 
- takes a cacheItem object and puts it in the hashMap
     
delete(key) 
- deletes the cacheItem object associated with the given key from the hashMap.

deleteLRU(requiredSpace) 
- deletes as many LRU objects as required to free up at least requiredSpace.


CacheItem
The data in the hashMap is stored in the form of cacheItem objects. We need to have
a cacheItem object to store in the map and cacheItem is what we get from the map
when we do a get. The object supports following methods. These objects are
read only.

getKey()
- returns the key associated with the cacheItem

getKeySize()
- returns the number of bytes in the key

getExpiryTime()
- returns the time in seconds since epoch when the current item will expire.

getFlags()
- returns the flags associated with the cacheItem

getDataSize()
- returns the size of the data stored in this cacheItem

getData()
- returns the data stored in the cacheItem as lua string. This is normally not
used because lua script don't need access to actual data unless using virtual keys.

delete()
- deletes the reference to the cacheItem object. Important to call this after get
from hashMap, after using the cacheItem.


Command
This represents the request coming from the client. This can be modified if
required by the script.

getCommand()
- returns the memcached the command being used.
  One of get, add, set, replace, prepend, append, cas, incr, decr, gets, delete,
  stats, flush_all, version, quit or verbosity
     
getKey()
- returns the key used in the command. This is nil in case of multi-get and other
  commands which don't have key as argument.

setKey(newKey)
- set the given string newKey as the key for the command object.

getKeySize()
- returns the size in bytes for the key

getExpiryTime()
- returns the expiryTime specified in the command.

setExpiryTime(newTime)
- sets the expiry time to new value newTime.

getFlags()
- returns the flags specified in the command. It is also used by the verbosity
  command to return the logging level.
 
setFlags(newFlags)
- sets the falgs in the command object to new flags.

getDelta()
- returns the delta value used in incr/decr commands

setDelta(newDelta)
- sets the delta value to new delta value

getNoReply()
- returns the boolean noReply from the command.

setNoReply(newNoReply)
- sets the value for the noReply

getDataSize()
- returns the size of data in the command, in bytes. This works for set, add, etc.

setData(newData)
- replaces the data in the command object with the newData string

getData()
- returns the data in the command object as lua string

newCacheItem()
- creates and returns a cacheItem object using the data in the command object.
  This is used with set/add/etc

writeCacheItem(cacheItem)
- writes the cacheItem in the memcached response format on the client socket.
   "VALUE key flags size\r\ndata\r\n"
  cas is not supported.

writeString(dataString)
- writes arbitrary string to the client socket. This is useful for writing
  "END\r\n", "STORED" and other standard memcached response strings.

hasMultipleKeys()
- returns the number of multi-get keys in the command

getMultipleKeys()
- returns a table of all the keys in the multi-get command

Apart from these, few helper methods defined in config.lua are also useful.

writeStringAsValue(command, key, value) 
- writes an arbitrary string in the "VALUE key flags size\r\ndata\r\n" format.
  flags is 0 and size is calculated using string.len


executeReadOnly(command, originalKey, objectType, cacheKey, func, ...) 
- helper function for read only operations on lua tables stored in the cache.
 It retrievs the data from the cache, uses table.unmarshal to construct the
 lua table and calls the function func with this lua table as the first
 argument. Any extra args passed to this function are passed to function func.


executeReadWrite(command, originalKey, objectType, cacheKey, func, ...) 
- helper function for read/write operations on lua tables stored in the cache
  orginal object is deleted and a new object is created with the same key with
  latest data values
 

executeNew(command, originalKey, objectType, cacheKey, func, ...)
- helper function for creation of new objects based on lua tables. if the key
  is in use, it is deleted before creating the new object

See set.lua, map.lua, quota.lua and swcounter.lua for example usage.

Cacheismo sliding window counter

Sliding window counters are useful for maintaining information about some last n transactions. One example could be a site like pingdom storing last n ping times for a site. Cacheismo comes with a default sliding window counter example which can be modified to suit your needs.  The current implementation supports new, add, getlast, getmin, getmax, getavg.

Note that these are memached get requests with special keys encoding our intent.  The keys follow the following syntax:
<ObjectType>:<Method>:<ObjectKey>:<Arg1>:<Arg2>

Object type is "swcounter". Cacheismo has other object types as well and more can be added by adding new lua scripts. Methods for swcounter object are "new", "add", "getmin", "getmax","getavg", "getlast", Object key we are using is "myCounter". You would need key per counter. Number of arguments would vary depending upon the method. Here new takes 1 arg (number of values to store), add takes 1 arg (value to store). Rest all methods take no arguments.

get swcounter:new:myCounter:2
VALUE  swcounter:myCounter:100 0 8 
CREATED 
-- Create a new counter which stores the last 2 values
get swcounter:add:myCounter:100
VALUE  swcounter:add:myCounter:100 0 7 
SUCCESS 
--store 100 as first value 
get swcounter:add:myCounter:50
VALUE  swcounter:add:myCounter:50 0 7 
SUCCESS 

-- store 50 as second value

get swcounter:getmin:myCounter
VALUE  swcounter:getmin:myCounter 0 2
50
-- get the min entry among all the stored values
get swcounter:getmax:myCounter
VALUE  swcounter:getmax:myCounter 0 3
100
-- get the max entry among all the stored values
get swcounter:getavg:myCounter
VALUE  swcounter:getavg:myCounter 0 2
75
-- get the average value for all the stored values

The complete implementation in lua looks like this.

====================================================

-- sliding window counter 

local function new(windowsize) 

  local  object = {}

  if (windowsize == nil) then

     windowsize = "16"

  end

  object.size   = tonumber(windowsize)

  object.index  = 0

  object.values = {}

  local count   = 0

  for count = 1,object.size do 

object.values[count] = 0

  end

  return object

end

local function add(object, value) 

   object.index = object.index+1

   if (object.index > object.size) then 

       object.index = 1 

   end 

   object.values[object.index] = tonumber (value) 

end

local function getlast(object)

   if (object.index ~= 0) then

       return object.values[object.index]

   end

   return 0

end

local function getmin(object)

   local count = 0

   local min   = math.huge

   for count = 1,object.size do

       if (object.values[count] < min) then

          min = object.values[count]

       end 

   end

   return min

end

local function getmax(object)

   local count = 0

   local max   = 0

   for count = 1,object.size do

       if (object.values[count] > max) then

          max = object.values[count]

       end 

   end

   return max

end

local function getavg(object)

   local count = 0

   local total   = 0

   for count = 1, object.size do

        total = total + object.values[count]

   end

   return (total/object.size)

end

=============================================================

Rate Limiting/Quota with Cacheismo

Cacheismo is a non-blocking single threaded high performance scriptable memcache. Instead of storing bytes or blobs in the cache, it can store objects defined in lua. No need to get the object, modify it and then save it back (with retry).  Directly modify object state in the cache itself by invoking methods of your objects. Because it is single threaded, modifications are automatically atomic. It is best suited for managing volatile application state at super speed (70K requests per second per core). Memcached would be more useful if your objects don't change much. But if they are changed frequently, memcached's update mechanism using retries would give you significantly low performance.

It comes with a default implementation of a quota object. You can change it as it suits you without touching the c code.  Here is how to use it.

Note that these are memached get requests with special keys encoding our intent.  The keys follow the following syntax:
<ObjectType>:<Method>:<ObjectKey>:<Arg1>:<Arg2>

Object type is "quota". Cacheismo has other object types as well and more can be added by adding new lua scripts. Methods for quota object are "new", "addandcheck", "reset". Object key we are using is "userKey". You would need key per user. Number of arguments would vary depending upon the method. Here new takes 2 args, addandcheck takes 1 and reset takes 0.

get quota:new:userKey:10:hour 
VALUE quota:new:userKey:10:hour 0 7 
CREATED
-- Creates a new quota object in cacheismo, that enforces 10 requests per hour.
get quota:addandcheck:userKey:10
VALUE quota:addandcheck:userKey:10 0 8
QUOTA_OK
-- Checking if user can use 10 units of quota. This succeeds. It is easy to create a default user object if it doesn't exist in this call also, instead of forcing users to always do a new first.
get quota:addandcheck:userKey:1
VALUE quota:addandcheck:userKey:1 0 14
QUOTA_EXCEEDED
-- Checking if user can use 1 more unit of quota. This fails as user exhausted it in previous call.
get quota:reset:userKey
VALUE quota:reset:userKey 0 8
QUOTA_OK
-- Reset the quota back to 0.


Cacheismo can do around 70K such operations per second running on single core. Like memcached, if single instance is not enough, multiple instances can be used. Cacheismo supports memcached ascii protocol. So any memcached client can be used to access it.

Introduction to cacheismo

=========================================================
-- simple quota implementation 
local quotaTypesMap = { month = 1, day = 1, hour = 1 }


local function findExpiresAt(quotaType)
       local current = os.date("*t")
       if (quotaType == "month") then
           current.month = current.month + 1
           current.day   = 1
           current.hour  = 0
           current.min   = 0
           current.sec   = 1
       elseif (quotaType == "day") then
           current.day   = current.day + 1
           current.hour  = 0
           current.min   = 0
           current.sec   = 1
       else 
           current.hour = current.hour+1
           current.min   = 0
           current.sec   = 1
       end 
    return os.time(current)
end


local function new(limit, quotaType) 
  if (quotaTypesMap[quotaType] == nil) then
      return nil
  end
  local  object    = {}
  object.limit     = tonumber(limit)
  object.type      = quotaType
  object.count     = 0
  object.expiresAt = findExpiresAt(quotaType)
  return object
end


local function addandcheck(object, value)
   local toUseValue = 1 
   if (value ~= nil) then
      toUseValue = tonumber(value)
   end
   if (os.time() > object.expiresAt) then
       object.count = 0;
       object.expiresAt = findExpiresAt(object.type) 
   end
   if ((object.count + toUseValue) > object.limit) then
       return -1
   else 
       object.count = object.count + toUseValue
       return 0
   end
end


local function reset(object) 
   if (os.time() > object.expiresAt) then
       object.count = 0
       object.expiresAt = findExpiresAt(object.type) 
   end
   object.count = 0
   return 0
end
===================================================================

Cacheismo Memory Allocation

Cacheismo uses two different kind of memory allocators. One is called chunkpool which is used to store the items and another is called fallocator which is used for request processing. Both have different design goals. chunkpool is designed to be reasonably fast but very compact. It tries to minimize memory wastage. fallocator on the other hand doesn't care about memory wastage but tries to be very fast. Here are the details.

fallocator 
fallocator is used when we are allocating memory temporarily and it is certain that it will be freed very soon. For example when reading data from socket and parsing the request, many small strings are generated but these are released as soon as the request is over, which could be a matter of micro seconds.   Cacheismo creates a fallocator per connection. fallocator not only allocates memory but also keeps track of memory. Thus if because of bad code we miss freeing some memory, it will be reclaimed when the fallocator is destroyed, essentially when the connection is closed.

fallocator internally allocates a big buffer (4KB) and keeps track of how much of the buffer is used. Malloc simply returns buffer+used at the pointer and increments used by size of malloc.  If the current buffer cannot handle the request, new buffer is allocated and memory given from new buffer. Used buffers are kept in a linked list. On free, fallocator decreases the refcount on the buffer which was used to allocate that memory. If the refcount becomes zero, the buffer is freed from the used buffers list. The benefit of this refcount based memory management is that it keeps memory usage to minimum. So if you are doing some malloc/frees in a loop, they will not exhaust the memory because refcount mechanism will ensure that they are released. If the malloc size is bigger than 4KB, fallocator defaults to malloc but keeps track of the buffer.

Moreover the 4KB buffer are cached to avoid going to malloc every time. This scheme of things give us O(1) malloc and  O(1) free in normal case (when  not going to system malloc).   It avoids fragmentation of the memory but could result in more memory being used than required at any point in time. It also gives us some protection against memory leaks.

chunkpool
chunkpool is used for storing the cacheitems and the goal here is to maximize memory utilization. chunkpool imposes two constrains on the application. One, it never allocates more than 4KB of contiguous memory. That is, if the application needs 8KB of memory, it can at max only get 2 4KB blocks and not a single 8KB block. Applications needs to be designed around this limitation. The second one is not a constraint, but an advice.  Apart from malloc, chunkpool exposes another function called relaxedMalloc which can return memory smaller than what application asked for. Thus if application asks for 4KB of memory, it might return 256 bytes or 1KB if it doesn't have a 4KB block. If the returned memory is too small, application can free it and deal with out of memory condition or if possible use large number of small buffers instead of small number of large buffers.

Because of these restrictions, fragmentation of memory doesn't makes the system unusable, but only adds  more memory management overhead. Internally we have slabs of size 16 + n16 upto 4096 bytes. Initially all memory is allocated to the 4KB slab. If applications asks for say 1KB of memory we split the 4Kb buffer into two buffer on of 3Kb and another of 1Kb. The 1 KB buffer goes to the application and the 3KB is stored in the 3Kb slab. Cacheismo chunkpool uses skiplist to keep track of non empty slabs. At runtime if we found that the slab of asked malloc size is empty, it quickly find the next not empty slab, splits the buffer and returns memory to the application. On free, chunkpool tries to merge the free buffers to create bigger buffers.  One limitation that I will fix later is that chunkpool memory uses the first four bytes for memory management. The implication is that any free buffer can only merge with logical next buffer and not logical previous buffer. This would require storing buffer size also at the end of the buffer and not just at the start.  To circumvent this problem, chunkpool requires an extra GC call where it tries to merge 8MB of pages at a time.  This is called once every second.  This is triggered only if at least 12% of memory is free and average chunk size is less than 256 bytes other wise it is a no-op. Even though chunkpool is capable of addressing upto 64GB of memory, it is advisable not to go beyond say 8GB.

chunkpool is a very efficient allocation manager for cache data, when compared with slab allocator.  It saves user the trouble of first thinking about what slab sizes should be used and maintains high HIT Rate when object sizes are changed because of changes in application code.  Cacheismo is not faster than memcached when comparing raw TPS, but when you take HIT Rate into account, it saves you lots of db accesses and increases overall application performance.  You can download it from here.

Cacheismo Cluster Update

I have started work on cacheismo cluster. I should have something working in next couple of days. The interface for lua scripts to get values from the cluster would need small change. Instead of only specifying the key, the caller will need to specify both dataKey as well as the virtualKey when looking up for a value. Cacheismo will use the dataKey to decide the server to which the virtualKey would be sent so that we don't end up in a endless loop. For dataKey's lookup is anyways simpler.

The cacheismo cluster specification (set of servers in the cluster) would be exposed via special virtual keys, so that it can be updated just like other objects. Each of the servers will monitor all other servers in the cluster and update its consistent hashing table so that requests are routed appropriately. This places a restriction on the number of servers we can add in a cluster. Maybe at some point what we will need is neste or multi-level consistent hashing where a special node will stand for multiple other nodes, something like a router which connects one network with another.

UPDATE: Cacheismo Cluster Update 2

Cacheismo Next Steps

Most important limitation of current approach is that cacheismo cannot be used in clustered mode. The reason being virtual key hashing might choose a server different from where the actual key is stored.
hash("set:new:myKey") is not equal to hash("set$myKey") is not equal to hash("set:count:myKey")

I did some experiments with lua threads and they look pretty cheap to create. The current plan is to make cacheismo nodes aware of each other and add client side support to cacheismo server. Instead of just looking up the key in local server memory, cacheismo will also do the hashing at server side on the actual key and fetch the results from the responsible cacheismo server.  This would double the latency but ensure that any virtual key can be thrown at any server and it will fetch the results.

For example consider something like "set:intersection:mySet1:mySet2:mySet3".
Lets assume that  mySet1, mySet2 and mySet3 are stored on different servers.
Trivial implementation would be to get mySet1, mySet2 and mySet3 to server executing the request and do the processing. Another possibility is to get sizes of mySet1, mySet2 and mySet3 and execute two parallel "set:intersection:smallest:other1" and "set:intersection:smallest:other2", followed by local intersection over the result.

Since cacheismo is scriptable, it is always possible to change this logic.  I will hide the non-blocking details using lua co-routines so writing such objects will not require "knowing" the asynchronous IO aspects, just some friendly function call like getValueForKey.

It might make it possible to do some in memory map-reduce with recursive calls for virtual keys running parallel stacks spread over the network, executing at the same time.

UPDATE: Cacheismo Cluster Update

Introducing Cacheismo

Cacheismo is scriptable object cache which can be used as a replacement for memcached. It supports memcache protocol (tcp and ascii only). It   automatically manages slab sizes, no user configuration is required.    Slabs are intelligently managed and they reconfigure themselves as per  the usage. PUT never ever fails. Details follow:     
- Efficient use of memory 

  Slab allocator used in memcached would waste memory if slab sizes
  and object sizes are not in sync. For example when using slab sizes
  of 64/128/256/512/1024 bytes object with size 513 is stored in 1024
  bytes slab, wasting almost half the memory. Cacheismo tries to 
  maintain memory used by object as close as possible to the object size

- Reclaimable Memory 

  Every byte is reclaimable in cacheismo. PUT never fails. You will
  never have slab full issues. More on memory management.

- True LRU support. Dependable cache

  Memcached will evict items from the slab in which object will be        placed. Cacheismo  will always use the LRU, irrespective of the size.   This ensures that the most  important data is always in the cache. No   surprises. If something is not in the cache, it either expired or was   least used.

- Server Side Scriptable objects

  The best part of using cacheismo is that is it fully scriptable
  in lua. Instead of cluttering your code with workarounds for 
  memcached read/write semantics, you can move that logic inside 
  cacheismo. 
  See Use cases for cacheismo 
      Rate Limiting Quota With Cacheismo
      Using Cacheismo Cluster For Real-time Analytics

  Sample objects map, set, quota and sliding window counters
  written in lua can be found in scripts directory. You could create 
  your own objects and place them in the scripts directory. 
  
  The interface for accessing scriptable objects is implemented via
  memcached get requests. For example: 
  get set:new:mykey     
  would create a new set object refered via myKey
  get set:put:myKey:a   
  would put key a in the set myKey
  get set:count:myKey   
  would return number of elements in the set
  get set:union:myKey1:myKey2   
  would return union of sets myKey1 and myKey2
  
  See scripts/set.lua for other functions.  
  
  I call this approach virtual keys. Reason for using it ...
  You can use existing memcached libraries to access cacheismo. 
  
Building - cacheismo depends on libevent and lua (luagit can 
also be used). Complete implementation is in about 5000 lines 
of code and fairly organised. 



Blue line is (memcachedTPS-cacheismoTPS)/memcachedTPS. This           essentially measures how much faster is memcached compared to         cacheismo. As you can see at max it is 20% faster.

Black line is (cacheismoHitRATE-memcachedHitRATE)/cacheismoHitRATE.   This essentially means how much memory efficient is cacheismo compared to memcached. For the first half of the test case we don't see any   difference. But in the second half cacheismo is at max about 80%      efficient. The reason for this behavior is that memcached can't       reassign slabs once they are used whereas cacheismo continues to      adjust with changing object sizes. Over time memcached commits memory to some slabs and hence its hit rate decreases over time.

Get source from GitHub

Google Group

UPDATE: Cacheismo Next Steps

What to do when your code suddenly slows down by 60%?

before running the tests multiple times to confirm that this indeed is the case...
before looking into the recent changes...
before looking into the test code...
before checking if swap is being used...
before checking if logging is enabled...
before doubting the kernel upgrade...
before using strace -c to find out unusual system calls...
before using valgrind --tool=callgrind to profile the code...

The first thing to do is to check cat /proc/cpuinfo and check in cpu Mz is same as advertised cpu speed. Laptops usually run in power saving mode. Not sure what is wrong with 64 bit Ubuntu 11.04, running on Lenovo Thinkpad T410. Even with power plugged in, it decided to run at 933Mz instead of 2.53 Gz.
Thanks to google for this link.

I am not with Anna

Call me pessimist, but I don't see corruption going away any time soon.

Loose definition of corruption would be using your "position" for "personal gains" instead of doing what is "right".  The operating word here is "position". How much corrupt can a beggar be? How much corrupt can a government be? Clearly government holds such an important position that it is very likely to be corrupt.  It pays to be corrupt government.

How about Lokpal? Does Lokpal holds important position? Yes it does and in some sense it holds a position more powerful than the government itself. Does it pays to be a corrupt Lokpal? Handsomely. Almost as much as a corrupt government.  I kind of believe in Murphy's Law as much as I believe in gravity.

Lokpal is as susceptible to corruption as much is government or a government employee. The process of choosing Lokpal is similar to aristocracy.  The difference is instead of choosing the government itself we are using it to choose the Lokpal. In some sense what Anna is saying is that Aristocracy is a better form of governance than democracy.  Is it?

Solving the corruption problem needs fundamental understanding and play between the concepts of "position", "personal gains" and what is "right".

Position:  Power corrupts, absolute power corrupts absolutely. How can we change the power equation?  Democracy gives people the power to change the government every five years. Government gets absolute power for five years. Here is my take on better democracy Dynamic Government.

Personal Gains:  This is usually money but it could be anything that can lead to money in the future. Again it need not be personal. It could easily go to friends and family.  I believe that the concept of money is broken. See Reinventing Money.  India badly needs to have Inheritance Tax.  In some sense money is equivalent to "position" if the person in "position" is corrupt. Hence absolute money for long duration is equivalent to absolute power and will lead to corruption. See  If Money Expired And People Had Choice .

What is Right: This one is hard to crack. Moral is right according to religion. Legal is right according to Law. Rational is right according to reason - but who knows what we are missing and how logical our minds think. Religion and Law are created by mortals only and could be wrong at times. I don't know the answers here but this is the direction I am looking at - On Reality Truth And Levels Of Abstractions and The Idiot Religion

I think corruption is a design issue, design of human interactions, how they help each other to survive together.

Mini Speed Breaker

The idea behind speed breaker is breaking the speed. The effectiveness of speed breaker drops with visibility. It is useless if driver cannot see it. In fact it can do more harm instead of providing safety.

Twist is, use a mini speed breaker before the actual speed breaker to give indication to the driver for the coming speed breaker. It is better than painting speed breakers or putting up boards because it doesn't needs  visibility for communication.  Basically if you hit a mini speed breaker, just reduce your speed as a big one will follow. 

Open Source vs Closed Source

Love this ;)

Optimizing time in getting off the plane

Two cases:  With or without check-in baggage?

• Get your hand baggage as soon as possible and stand in the line.
• This will take around 5 minutes.
• Get into the bus, but make sure you take the seat close to the gate or preferably stand
• Get out of the bus and go out (no check-in baggage) 
• Get out of the bus and wait for baggage to arrive. 
• Usually baggage will come first if you were the last one to check-in the flight (FILO).
• Wait for baggage 

When getting out of the landing airport is decided by when you arrive at the departure airport, then why do people continue to fight for being first to get out of the plane. 

Why Low Scoring IPL Matches Last Long

IPL makes money via advertisements in the match.  If the match finishes in 10 overs instead of 20, it means loss of revenue. It seems like  no matter what, both "strategic timeouts" have to happen in any match. I am not aware of any match in which a team won before the "strategic timeout" irrespective of the required runs.

Sometimes I wonder if sports is a form of Reality TV.

Named Queries in Relational Databases

Databases don't provide a way to create named queries with runtime arguments.  Yes we have views but most databases are accessed via programming languages where we need to construct the SQL query. We do have the Prepared statements but again they are programming language constructs and not named database level constructs.

Benefits:
1) No need to parse queries. They are callable by name.
2) Prevents SQL injection attacks because of not parsing.
3) Easy integration with programming languages.  db.execute("myQuery", arg1, arg2,...)

4) Instead of trying to optimize every SQL query, every time, database will have the list of the most important queries it needs to answer. This allows databases to save time by pre-planning the execution in the short run and in the long run, databases can optimize the layout of files or creating of indexes automatically to optimize the access pattern.

5) Easy to change db logic. As long as the query returns the same fields (may be more), you can easily change the actual query logic without changing application code.

Their is a lot in the name. 

The better side of corruption

Corruption is an enabler. It thrives on inefficiency, shortcomings of the system and provides a way to get things done for a price.  People pay for stuff that is valuable to them, hence at the very least corruption is a great validation...it validates, confirms that something is important.  The answer to the question, how do I make something better becomes self evident...fix things where their is corruption.

I wanted to move from Reliance CDMA to Airtel GSM. After 2 weeks, 6 visits of 30 minutes each and asking 10 times when should I disconnect my Reliance connection, the ultimate answer is - "Sorry your request couldn't be processed because you have not disconnected your Reliance connection".

I wish Airtel guys were corrupt. They would have known what people want, how much they want it and  the best interface to execute it.

Customer Support

Customers really don't want Customer Support, they just want working products. 

Price & Retailer Margin

Assuming price is determined by the demand and supply rules, how does one decides the retailer margin?

Let P be the price and M% be the retailers margin and C the expected number of units sold per month, then

Earnings per month for retailer = P * M * C
This number should be high enough to make a living for the retailer after deducting the cost of running the business.

Examples:

• Mobile prices are high and number of units sold per month is good, causing margins to be low
• Taps, shower, lights, basins, etc are sold at low volumes, prices are OK and margins are very high
• FMCG products have very low margins
• Kitchen chimneys and burners are high priced, volumes are low and margins are high
• Cigarettes have high volume and hence margins are low
• High end cosmetics have very low volumes, hence has reasonable margins
• Books have low volume, unit price is OK, margins should be high
• Text Books have high volume, unit price is OK, but they are sold only once in a year, very much like crackers on diwali.
• Pet food also has low volume and hence high margins
• Furniture again low volume, high price and hence margins ought to be reasonable.

I guess the opportunity for e-commerce is creating business efficiency for low volume products to drive their prices down by passing on the profits to customers making it impossible for some of the retailers to justify their business. The high volume products will always have lower margins and hence price is not what will drive buying on the net, but the other factors like trust, service, range, etc. It also opens up the market for products which couldn't find retail space because they couldn't fulfill this equation.

Another side effect of this equation is what will be counter intuitive for most economists. Under specific circumstances, when the demand goes low, prices don't really go down, they increase. The reason - retailer needs to meet his monthly needs. If he makes less money, the only way to go back to original income levels is increasing the price and not decreasing the price. The specific circumstances are:

• All retailers are friends (taxi operators, unions, etc) 
• Customer lifetime values is same as that of current transaction. For example a tourist hiring a taxi is not going to remember or call the same guy on his next visit.
• How much time does customer has to make the decision. The lesser the time, less choice he has.
• Customer already committed part of the money. 
• Does competition exists? 
• Is trust part of the equation?  

Old Poem

मुर्गी से आया है अंडा

या अंडे से मुर्गी आई है 

कविता ने कवी को बनाया है 

या कवी ने कविता बनाई है

दुविधा में हूँ फंसा हुआ 

क्या सच है क्या सचायी है

क्या सच की कोई परिभासा है 

क्या सच झूठ के बीच कोई रेखा है 

जो देखा है वोह सच है क्या 

क्या सच को किसी ने देखा है 

क्या वही है सच जो दीखता है 

या वोह जो आप समझते हैं

लकिन जो आप समझते है 

क्या आप उसे समझते हैं

मैंने सच को समझा नहीं 

जो समझा है समझा दे मुझे 

मैं उलझा हूँ उलझन मैं मेरी 

मुजको मुझेसे मिला दो हरे

Mean And/Or Stupid

Mean and Stupid are two labels which can be used to describe people. Just like tall, dark, handsome or generous, wealthy, shy, happy, content, etc. All labels are dangerous, but these two in-particular are particularly dangerous.

Label someone mean and it gives us the license to be mean to them.
Label someone stupid and it gives us the license to not listen to them. Someone who doesn't appreciates something intelligent said to them is by definition stupid.

Thinking someone is mean and/or stupid eventually makes us  mean and stupid and being mean and stupid anyway causes other people to be mean and stupid. It is a vicious circle.  It is fiction creating hardcore reality which is actually factitious.

Just think how many people are mean and stupid just because we think they are. You are what you think you are is alright, but you are what you think someone else is, is scary.