セッションストレージとリーダーボードのパターン

サーバー側セッションの問題

セッションストレージとしてのRedis

# Session storage with Redis (Python Flask example)
import redis, json, uuid
from datetime import timedelta

redis_client = redis.Redis(host='prod-redis-primary', port=6379)
SESSION_TTL = int(timedelta(hours=8).total_seconds())

def create_session(user_id):
    session_id = str(uuid.uuid4())
    session_data = {'user_id': user_id, 'logged_in': True}
    redis_client.setex(f'session:{session_id}', SESSION_TTL, json.dumps(session_data))
    return session_id

def get_session(session_id):
    data = redis_client.get(f'session:{session_id}')
    return json.loads(data) if data else None

セッションTTLとスライディング有効期限

# Sliding TTL session implementation
def get_session_with_sliding_ttl(session_id, redis_client, ttl_seconds=1800):
    session_key = f'session:{session_id}'

    # Pipeline: GET + EXPIRE in one round trip
    pipe = redis_client.pipeline()
    pipe.get(session_key)
    pipe.expire(session_key, ttl_seconds)  # Reset TTL on access
    results = pipe.execute()

    data = results[0]
    if data:
        return json.loads(data)
    return None   # Session expired or not found

Redisでのショッピングカート

# Shopping cart operations using Redis Hash
cart_key = f'cart:{user_id}'

# Add item (or increase quantity)
# HINCRBY cart:user42 SKU-001 2
redis_client.hincrby(cart_key, 'SKU-001', 2)

# Remove item
# HDEL cart:user42 SKU-001
redis_client.hdel(cart_key, 'SKU-001')

# Get all items in cart
# HGETALL cart:user42
cart = redis_client.hgetall(cart_key)  # {b'SKU-001': b'2', b'SKU-002': b'1'}

# Set TTL for cart abandonment (24 hours)
redis_client.expire(cart_key, 86400)

リーダーボードのアーキテクチャ

# Real-time leaderboard with Redis Sorted Set

# Add or update a player's score
# ZADD game:weekly:leaderboard 15750 'player:alice'
redis_client.zadd('game:weekly:leaderboard', {'player:alice': 15750})

# Increment score (atomic)
# ZINCRBY game:weekly:leaderboard 500 'player:alice'
redis_client.zincrby('game:weekly:leaderboard', 500, 'player:alice')

# Get top 10 players (highest scores first)
# ZREVRANGE game:weekly:leaderboard 0 9 WITHSCORES
top_10 = redis_client.zrevrange('game:weekly:leaderboard', 0, 9, withscores=True)

プレイヤーの順位と周辺のプレイヤー

# Get Alice's rank (0-indexed, so add 1 for display)
# ZREVRANK game:weekly:leaderboard 'player:alice'
rank = redis_client.zrevrank('game:weekly:leaderboard', 'player:alice')
print(f'Alice is rank #{rank + 1}')

# Get 5 players above and below Alice
start = max(0, rank - 5)
end = rank + 5
nearby = redis_client.zrevrange(
    'game:weekly:leaderboard', start, end, withscores=True
)
print('Players near Alice:', nearby)

Redisによるレート制限

# Sliding window rate limiter (100 requests per 60 seconds)
import time

def is_rate_limited(user_id, redis_client, limit=100, window_seconds=60):
    key = f'ratelimit:{user_id}'
    now = time.time()
    window_start = now - window_seconds

    pipe = redis_client.pipeline()
    pipe.zremrangebyscore(key, '-inf', window_start)  # Remove old
    pipe.zcard(key)                                    # Count current
    pipe.zadd(key, {str(now): now})                   # Add this request
    pipe.expire(key, window_seconds)
    results = pipe.execute()

    request_count = results[1]
    return request_count >= limit  # True = rate limited

Redisによる分散ロック

# Distributed lock with Redis SET NX EX
import uuid

def acquire_lock(redis_client, resource, ttl_seconds=30):
    lock_id = str(uuid.uuid4())  # Unique ID to identify this lock holder
    key = f'lock:{resource}'
    acquired = redis_client.set(key, lock_id, nx=True, ex=ttl_seconds)
    return lock_id if acquired else None

def release_lock(redis_client, resource, lock_id):
    key = f'lock:{resource}'
    # Only delete if we still own the lock (Lua script for atomicity)
    lua = 'if redis.call("get",KEYS[1])==ARGV[1] then return redis.call("del",KEYS[1]) else return 0 end'
    redis_client.eval(lua, 1, key, lock_id)

セッションストレージ：ElastiCacheとDynamoDBの比較

Redisによる地理空間インデックス

# Store driver locations
# GEOADD drivers 13.361389 38.115556 'driver:001'
# GEOADD drivers 15.087269 37.502669 'driver:002'

# Find all drivers within 10 km of a point
# GEOSEARCH drivers FROMLONLAT 13.5 38.1 BYRADIUS 10 km ASC COUNT 5 WITHCOORD

# Result: sorted list of driver IDs within 10 km with coordinates

一意な訪問者数のカウントに使うHyperLogLog

# Count unique daily visitors using HyperLogLog
date = '2024-01-15'
hll_key = f'unique_visitors:{date}'

# Track a visitor (PFADD is idempotent for the same user)
# PFADD unique_visitors:2024-01-15 'user:12345'
redis_client.pfadd(hll_key, 'user:12345')
redis_client.pfadd(hll_key, 'user:67890')
redis_client.pfadd(hll_key, 'user:12345')  # Duplicate — not counted again

# Get estimated unique visitor count
# PFCOUNT unique_visitors:2024-01-15
count = redis_client.pfcount(hll_key)
print(f'Unique visitors today (estimate): {count}')

クイックチェック

レッスンのまとめ