kailash 0.8.3__py3-none-any.whl → 0.8.5__py3-none-any.whl

kailash/edge/resource/cost_optimizer.py

@@ -0,0 +1,949 @@
+"""Cost optimizer for intelligent edge resource cost management.
+
+This module provides multi-cloud cost optimization, spot instance management,
+reserved capacity planning, and ROI-based allocation decisions.
+"""
+
+import asyncio
+import json
+import logging
+from collections import defaultdict
+from dataclasses import dataclass, field
+from datetime import datetime, timedelta
+from enum import Enum
+from typing import Any, Dict, List, Optional, Tuple
+
+import numpy as np
+
+
+class CloudProvider(Enum):
+    """Supported cloud providers."""
+
+    AWS = "aws"
+    GCP = "gcp"
+    AZURE = "azure"
+    ALIBABA = "alibaba"
+    EDGE_LOCAL = "edge_local"
+
+
+class InstanceType(Enum):
+    """Instance pricing types."""
+
+    ON_DEMAND = "on_demand"
+    SPOT = "spot"
+    RESERVED = "reserved"
+    SAVINGS_PLAN = "savings_plan"
+    DEDICATED = "dedicated"
+
+
+class OptimizationStrategy(Enum):
+    """Cost optimization strategies."""
+
+    MINIMIZE_COST = "minimize_cost"
+    BALANCE_COST_PERFORMANCE = "balance_cost_performance"
+    MAXIMIZE_PERFORMANCE = "maximize_performance"
+    PREDICTABLE_COST = "predictable_cost"
+    RISK_AVERSE = "risk_averse"
+
+
+@dataclass
+class CostMetric:
+    """Cost measurement for resources."""
+
+    timestamp: datetime
+    edge_node: str
+    resource_type: str
+    provider: CloudProvider
+    instance_type: InstanceType
+    cost_per_hour: float
+    usage_hours: float
+    total_cost: float
+    currency: str = "USD"
+    metadata: Dict[str, Any] = field(default_factory=dict)
+
+    def to_dict(self) -> Dict[str, Any]:
+        """Convert to dictionary."""
+        return {
+            "timestamp": self.timestamp.isoformat(),
+            "edge_node": self.edge_node,
+            "resource_type": self.resource_type,
+            "provider": self.provider.value,
+            "instance_type": self.instance_type.value,
+            "cost_per_hour": self.cost_per_hour,
+            "usage_hours": self.usage_hours,
+            "total_cost": self.total_cost,
+            "currency": self.currency,
+            "metadata": self.metadata,
+        }
+
+
+@dataclass
+class CostOptimization:
+    """Cost optimization recommendation."""
+
+    optimization_id: str
+    edge_node: str
+    current_setup: Dict[str, Any]
+    recommended_setup: Dict[str, Any]
+    estimated_savings: float
+    savings_percentage: float
+    confidence: float
+    implementation_effort: str  # low, medium, high
+    risk_level: str  # low, medium, high
+    reasoning: List[str] = field(default_factory=list)
+
+    def to_dict(self) -> Dict[str, Any]:
+        """Convert to dictionary."""
+        return {
+            "optimization_id": self.optimization_id,
+            "edge_node": self.edge_node,
+            "current_setup": self.current_setup,
+            "recommended_setup": self.recommended_setup,
+            "estimated_savings": self.estimated_savings,
+            "savings_percentage": self.savings_percentage,
+            "confidence": self.confidence,
+            "implementation_effort": self.implementation_effort,
+            "risk_level": self.risk_level,
+            "reasoning": self.reasoning,
+        }
+
+
+@dataclass
+class SpotInstanceRecommendation:
+    """Spot instance optimization recommendation."""
+
+    edge_node: str
+    current_on_demand_cost: float
+    spot_cost: float
+    potential_savings: float
+    interruption_risk: float
+    recommended_strategy: str
+    backup_plan: Dict[str, Any]
+
+    def to_dict(self) -> Dict[str, Any]:
+        """Convert to dictionary."""
+        return {
+            "edge_node": self.edge_node,
+            "current_on_demand_cost": self.current_on_demand_cost,
+            "spot_cost": self.spot_cost,
+            "potential_savings": self.potential_savings,
+            "savings_percentage": (
+                (self.potential_savings / self.current_on_demand_cost * 100)
+                if self.current_on_demand_cost > 0
+                else 0
+            ),
+            "interruption_risk": self.interruption_risk,
+            "recommended_strategy": self.recommended_strategy,
+            "backup_plan": self.backup_plan,
+        }
+
+
+@dataclass
+class ReservationRecommendation:
+    """Reserved capacity recommendation."""
+
+    resource_type: str
+    provider: CloudProvider
+    commitment_length: int  # months
+    upfront_cost: float
+    monthly_cost: float
+    on_demand_equivalent: float
+    total_savings: float
+    breakeven_months: int
+    utilization_requirement: float
+
+    def to_dict(self) -> Dict[str, Any]:
+        """Convert to dictionary."""
+        return {
+            "resource_type": self.resource_type,
+            "provider": self.provider.value,
+            "commitment_length": self.commitment_length,
+            "upfront_cost": self.upfront_cost,
+            "monthly_cost": self.monthly_cost,
+            "on_demand_equivalent": self.on_demand_equivalent,
+            "total_savings": self.total_savings,
+            "savings_percentage": (
+                (self.total_savings / self.on_demand_equivalent * 100)
+                if self.on_demand_equivalent > 0
+                else 0
+            ),
+            "breakeven_months": self.breakeven_months,
+            "utilization_requirement": self.utilization_requirement,
+        }
+
+
+class CostOptimizer:
+    """Multi-cloud cost optimizer for edge resources."""
+
+    def __init__(
+        self,
+        cost_history_days: int = 30,
+        optimization_interval: int = 3600,  # 1 hour
+        savings_threshold: float = 0.1,  # 10% minimum savings
+        risk_tolerance: str = "medium",
+    ):
+        """Initialize cost optimizer.
+
+        Args:
+            cost_history_days: Days of cost history to analyze
+            optimization_interval: How often to run optimization
+            savings_threshold: Minimum savings percentage to recommend
+            risk_tolerance: Risk tolerance (low, medium, high)
+        """
+        self.cost_history_days = cost_history_days
+        self.optimization_interval = optimization_interval
+        self.savings_threshold = savings_threshold
+        self.risk_tolerance = risk_tolerance
+
+        # Cost data storage
+        self.cost_metrics: List[CostMetric] = []
+        self.provider_pricing: Dict[str, Dict[str, Any]] = {}
+
+        # Optimization history
+        self.optimizations: List[CostOptimization] = []
+        self.implemented_optimizations: List[str] = []
+
+        # Background task
+        self._optimization_task: Optional[asyncio.Task] = None
+
+        self.logger = logging.getLogger(__name__)
+
+        # Initialize default pricing data
+        self._initialize_pricing_data()
+
+    async def start(self):
+        """Start background optimization."""
+        if not self._optimization_task:
+            self._optimization_task = asyncio.create_task(self._optimization_loop())
+            self.logger.info("Cost optimizer started")
+
+    async def stop(self):
+        """Stop background optimization."""
+        if self._optimization_task:
+            self._optimization_task.cancel()
+            try:
+                await self._optimization_task
+            except asyncio.CancelledError:
+                pass
+            self._optimization_task = None
+            self.logger.info("Cost optimizer stopped")
+
+    async def record_cost(self, cost_metric: CostMetric):
+        """Record a cost metric.
+
+        Args:
+            cost_metric: Cost metric to record
+        """
+        self.cost_metrics.append(cost_metric)
+
+        # Keep only recent history
+        cutoff = datetime.now() - timedelta(days=self.cost_history_days)
+        self.cost_metrics = [m for m in self.cost_metrics if m.timestamp > cutoff]
+
+    async def optimize_costs(
+        self,
+        strategy: OptimizationStrategy = OptimizationStrategy.BALANCE_COST_PERFORMANCE,
+        edge_nodes: Optional[List[str]] = None,
+    ) -> List[CostOptimization]:
+        """Generate cost optimization recommendations.
+
+        Args:
+            strategy: Optimization strategy
+            edge_nodes: Specific edge nodes to optimize
+
+        Returns:
+            List of cost optimizations
+        """
+        optimizations = []
+
+        # Get nodes to analyze
+        nodes_to_analyze = edge_nodes if edge_nodes else self._get_all_edge_nodes()
+
+        for node in nodes_to_analyze:
+            # Analyze current costs
+            current_costs = self._analyze_node_costs(node)
+
+            if not current_costs:
+                continue
+
+            # Generate optimization recommendations
+            node_optimizations = await self._optimize_node_costs(
+                node, current_costs, strategy
+            )
+
+            optimizations.extend(node_optimizations)
+
+        # Filter by savings threshold
+        significant_optimizations = [
+            opt
+            for opt in optimizations
+            if opt.savings_percentage >= self.savings_threshold * 100
+        ]
+
+        # Store optimizations
+        self.optimizations.extend(significant_optimizations)
+
+        return significant_optimizations
+
+    async def get_spot_recommendations(
+        self, edge_nodes: Optional[List[str]] = None
+    ) -> List[SpotInstanceRecommendation]:
+        """Get spot instance recommendations.
+
+        Args:
+            edge_nodes: Specific edge nodes to analyze
+
+        Returns:
+            List of spot instance recommendations
+        """
+        recommendations = []
+
+        nodes_to_analyze = edge_nodes if edge_nodes else self._get_all_edge_nodes()
+
+        for node in nodes_to_analyze:
+            current_costs = self._analyze_node_costs(node)
+
+            if not current_costs:
+                continue
+
+            # Check if spot instances would be beneficial
+            spot_rec = await self._analyze_spot_opportunity(node, current_costs)
+
+            if spot_rec and spot_rec.potential_savings > 0:
+                recommendations.append(spot_rec)
+
+        return recommendations
+
+    async def get_reservation_recommendations(
+        self, providers: Optional[List[CloudProvider]] = None
+    ) -> List[ReservationRecommendation]:
+        """Get reserved capacity recommendations.
+
+        Args:
+            providers: Specific providers to analyze
+
+        Returns:
+            List of reservation recommendations
+        """
+        recommendations = []
+
+        providers_to_analyze = providers if providers else list(CloudProvider)
+
+        for provider in providers_to_analyze:
+            # Analyze usage patterns for this provider
+            usage_patterns = self._analyze_provider_usage(provider)
+
+            for resource_type, usage in usage_patterns.items():
+                # Check if reservation would be beneficial
+                reservation_rec = await self._analyze_reservation_opportunity(
+                    provider, resource_type, usage
+                )
+
+                if reservation_rec and reservation_rec.total_savings > 0:
+                    recommendations.append(reservation_rec)
+
+        return recommendations
+
+    async def calculate_roi(
+        self, optimization: CostOptimization, implementation_cost: float = 0.0
+    ) -> Dict[str, Any]:
+        """Calculate ROI for an optimization.
+
+        Args:
+            optimization: Cost optimization to analyze
+            implementation_cost: One-time implementation cost
+
+        Returns:
+            ROI analysis
+        """
+        monthly_savings = optimization.estimated_savings
+
+        # Calculate payback period
+        payback_months = (
+            implementation_cost / monthly_savings
+            if monthly_savings > 0
+            else float("inf")
+        )
+
+        # Calculate 1-year ROI
+        annual_savings = monthly_savings * 12
+        roi_percentage = (
+            ((annual_savings - implementation_cost) / implementation_cost * 100)
+            if implementation_cost > 0
+            else float("inf")
+        )
+
+        # Risk-adjusted ROI
+        risk_multiplier = self._get_risk_multiplier(optimization.risk_level)
+        risk_adjusted_roi = roi_percentage * risk_multiplier
+
+        return {
+            "optimization_id": optimization.optimization_id,
+            "monthly_savings": monthly_savings,
+            "annual_savings": annual_savings,
+            "implementation_cost": implementation_cost,
+            "payback_months": payback_months,
+            "roi_percentage": roi_percentage,
+            "risk_adjusted_roi": risk_adjusted_roi,
+            "recommendation": self._get_roi_recommendation(
+                roi_percentage, payback_months
+            ),
+        }
+
+    async def get_cost_forecast(
+        self, forecast_months: int = 12, include_optimizations: bool = True
+    ) -> Dict[str, Any]:
+        """Get cost forecast with and without optimizations.
+
+        Args:
+            forecast_months: Months to forecast
+            include_optimizations: Include optimization impact
+
+        Returns:
+            Cost forecast
+        """
+        # Calculate current monthly spend
+        current_monthly = self._calculate_current_monthly_spend()
+
+        # Project baseline costs
+        baseline_forecast = []
+        optimized_forecast = []
+
+        for month in range(forecast_months):
+            # Apply growth assumptions
+            growth_factor = 1 + (0.05 * month / 12)  # 5% annual growth
+            baseline_cost = current_monthly * growth_factor
+            baseline_forecast.append(baseline_cost)
+
+            # Apply optimizations if requested
+            if include_optimizations:
+                total_savings = sum(opt.estimated_savings for opt in self.optimizations)
+                optimized_cost = baseline_cost - total_savings
+                optimized_forecast.append(max(0, optimized_cost))
+            else:
+                optimized_forecast.append(baseline_cost)
+
+        total_baseline = sum(baseline_forecast)
+        total_optimized = sum(optimized_forecast)
+        total_savings = total_baseline - total_optimized
+
+        return {
+            "forecast_months": forecast_months,
+            "current_monthly_spend": current_monthly,
+            "baseline_forecast": baseline_forecast,
+            "optimized_forecast": optimized_forecast if include_optimizations else None,
+            "total_baseline_cost": total_baseline,
+            "total_optimized_cost": total_optimized,
+            "total_projected_savings": total_savings,
+            "savings_percentage": (
+                (total_savings / total_baseline * 100) if total_baseline > 0 else 0
+            ),
+        }
+
+    async def _optimization_loop(self):
+        """Background optimization loop."""
+        while True:
+            try:
+                await asyncio.sleep(self.optimization_interval)
+
+                # Run automatic optimization
+                optimizations = await self.optimize_costs()
+
+                if optimizations:
+                    self.logger.info(
+                        f"Found {len(optimizations)} cost optimization opportunities"
+                    )
+
+            except asyncio.CancelledError:
+                break
+            except Exception as e:
+                self.logger.error(f"Optimization loop error: {e}")
+
+    def _initialize_pricing_data(self):
+        """Initialize default pricing data for providers."""
+        # Simplified pricing data - in production, this would come from APIs
+        self.provider_pricing = {
+            CloudProvider.AWS.value: {
+                "cpu": {
+                    "on_demand": 0.10,  # per vCPU hour
+                    "spot": 0.03,
+                    "reserved_1yr": 0.07,
+                    "reserved_3yr": 0.05,
+                },
+                "memory": {
+                    "on_demand": 0.01,  # per GB hour
+                    "spot": 0.003,
+                    "reserved_1yr": 0.007,
+                    "reserved_3yr": 0.005,
+                },
+                "storage": {"on_demand": 0.10, "reserved": 0.08},  # per GB month
+            },
+            CloudProvider.GCP.value: {
+                "cpu": {
+                    "on_demand": 0.09,
+                    "preemptible": 0.025,
+                    "committed_1yr": 0.065,
+                    "committed_3yr": 0.045,
+                },
+                "memory": {
+                    "on_demand": 0.009,
+                    "preemptible": 0.0025,
+                    "committed_1yr": 0.0065,
+                    "committed_3yr": 0.0045,
+                },
+            },
+            CloudProvider.AZURE.value: {
+                "cpu": {
+                    "on_demand": 0.11,
+                    "spot": 0.035,
+                    "reserved_1yr": 0.075,
+                    "reserved_3yr": 0.055,
+                },
+                "memory": {
+                    "on_demand": 0.011,
+                    "spot": 0.0035,
+                    "reserved_1yr": 0.0075,
+                    "reserved_3yr": 0.0055,
+                },
+            },
+        }
+
+    def _get_all_edge_nodes(self) -> List[str]:
+        """Get all edge nodes from cost metrics."""
+        return list(set(metric.edge_node for metric in self.cost_metrics))
+
+    def _analyze_node_costs(self, edge_node: str) -> Dict[str, Any]:
+        """Analyze costs for a specific edge node."""
+        node_metrics = [m for m in self.cost_metrics if m.edge_node == edge_node]
+
+        if not node_metrics:
+            return {}
+
+        # Group by resource type and instance type
+        cost_breakdown = defaultdict(lambda: defaultdict(float))
+
+        for metric in node_metrics:
+            key = f"{metric.resource_type}_{metric.instance_type.value}"
+            cost_breakdown[metric.resource_type][
+                metric.instance_type.value
+            ] += metric.total_cost
+
+        # Calculate totals
+        total_cost = sum(
+            sum(instance_costs.values()) for instance_costs in cost_breakdown.values()
+        )
+
+        return {
+            "edge_node": edge_node,
+            "total_cost": total_cost,
+            "cost_breakdown": dict(cost_breakdown),
+            "metrics_count": len(node_metrics),
+            "last_updated": max(m.timestamp for m in node_metrics),
+        }
+
+    async def _optimize_node_costs(
+        self,
+        edge_node: str,
+        current_costs: Dict[str, Any],
+        strategy: OptimizationStrategy,
+    ) -> List[CostOptimization]:
+        """Optimize costs for a specific node."""
+        optimizations = []
+
+        # Analyze each resource type
+        for resource_type, instance_costs in current_costs["cost_breakdown"].items():
+            # Check for spot instance opportunities
+            spot_opt = await self._check_spot_optimization(
+                edge_node, resource_type, instance_costs, strategy
+            )
+            if spot_opt:
+                optimizations.append(spot_opt)
+
+            # Check for reserved instance opportunities
+            reserved_opt = await self._check_reserved_optimization(
+                edge_node, resource_type, instance_costs, strategy
+            )
+            if reserved_opt:
+                optimizations.append(reserved_opt)
+
+            # Check for right-sizing opportunities
+            rightsizing_opt = await self._check_rightsizing_optimization(
+                edge_node, resource_type, instance_costs, strategy
+            )
+            if rightsizing_opt:
+                optimizations.append(rightsizing_opt)
+
+        return optimizations
+
+    async def _check_spot_optimization(
+        self,
+        edge_node: str,
+        resource_type: str,
+        instance_costs: Dict[str, float],
+        strategy: OptimizationStrategy,
+    ) -> Optional[CostOptimization]:
+        """Check for spot instance optimization opportunities."""
+        on_demand_cost = instance_costs.get("on_demand", 0)
+
+        if on_demand_cost == 0:
+            return None
+
+        # Get spot pricing
+        spot_cost = on_demand_cost * 0.3  # Assume 70% savings
+        potential_savings = on_demand_cost - spot_cost
+
+        # Check if savings meet threshold
+        savings_percentage = potential_savings / on_demand_cost * 100
+
+        if savings_percentage < self.savings_threshold * 100:
+            return None
+
+        # Assess risk based on workload characteristics
+        interruption_risk = self._assess_interruption_risk(edge_node, resource_type)
+
+        # Strategy-based risk tolerance
+        if strategy == OptimizationStrategy.RISK_AVERSE and interruption_risk > 0.3:
+            return None
+
+        if (
+            strategy == OptimizationStrategy.PREDICTABLE_COST
+            and interruption_risk > 0.1
+        ):
+            return None
+
+        return CostOptimization(
+            optimization_id=f"spot_{edge_node}_{resource_type}_{datetime.now().timestamp()}",
+            edge_node=edge_node,
+            current_setup={
+                "instance_type": "on_demand",
+                "cost": on_demand_cost,
+                "resource_type": resource_type,
+            },
+            recommended_setup={
+                "instance_type": "spot",
+                "cost": spot_cost,
+                "resource_type": resource_type,
+                "interruption_risk": interruption_risk,
+            },
+            estimated_savings=potential_savings,
+            savings_percentage=savings_percentage,
+            confidence=0.8,
+            implementation_effort="low",
+            risk_level="medium" if interruption_risk > 0.2 else "low",
+            reasoning=[
+                f"Spot instances offer {savings_percentage:.1f}% cost savings",
+                f"Interruption risk is {interruption_risk:.1%}",
+                "Workload appears suitable for spot instances",
+            ],
+        )
+
+    async def _check_reserved_optimization(
+        self,
+        edge_node: str,
+        resource_type: str,
+        instance_costs: Dict[str, float],
+        strategy: OptimizationStrategy,
+    ) -> Optional[CostOptimization]:
+        """Check for reserved instance optimization opportunities."""
+        on_demand_cost = instance_costs.get("on_demand", 0)
+
+        if on_demand_cost == 0:
+            return None
+
+        # Calculate usage consistency
+        usage_consistency = self._calculate_usage_consistency(edge_node, resource_type)
+
+        # Reserved instances only make sense for consistent usage
+        if usage_consistency < 0.7:
+            return None
+
+        # Calculate reserved cost (assume 30% savings for 1-year)
+        reserved_cost = on_demand_cost * 0.7
+        potential_savings = on_demand_cost - reserved_cost
+        savings_percentage = potential_savings / on_demand_cost * 100
+
+        if savings_percentage < self.savings_threshold * 100:
+            return None
+
+        return CostOptimization(
+            optimization_id=f"reserved_{edge_node}_{resource_type}_{datetime.now().timestamp()}",
+            edge_node=edge_node,
+            current_setup={
+                "instance_type": "on_demand",
+                "cost": on_demand_cost,
+                "resource_type": resource_type,
+            },
+            recommended_setup={
+                "instance_type": "reserved_1yr",
+                "cost": reserved_cost,
+                "resource_type": resource_type,
+                "commitment": "1 year",
+            },
+            estimated_savings=potential_savings,
+            savings_percentage=savings_percentage,
+            confidence=0.9,
+            implementation_effort="low",
+            risk_level="low",
+            reasoning=[
+                f"Reserved instances offer {savings_percentage:.1f}% cost savings",
+                f"Usage consistency is {usage_consistency:.1%}",
+                "1-year commitment recommended based on usage patterns",
+            ],
+        )
+
+    async def _check_rightsizing_optimization(
+        self,
+        edge_node: str,
+        resource_type: str,
+        instance_costs: Dict[str, float],
+        strategy: OptimizationStrategy,
+    ) -> Optional[CostOptimization]:
+        """Check for right-sizing optimization opportunities."""
+        # Analyze actual resource utilization
+        utilization = self._get_resource_utilization(edge_node, resource_type)
+
+        if utilization is None or utilization > 0.8:  # Well utilized
+            return None
+
+        # Calculate right-sized cost
+        utilization_factor = max(utilization * 1.2, 0.5)  # 20% buffer, minimum 50%
+        current_cost = sum(instance_costs.values())
+        rightsized_cost = current_cost * utilization_factor
+        potential_savings = current_cost - rightsized_cost
+        savings_percentage = potential_savings / current_cost * 100
+
+        if savings_percentage < self.savings_threshold * 100:
+            return None
+
+        # Risk assessment
+        risk_level = "low" if utilization < 0.5 else "medium"
+
+        return CostOptimization(
+            optimization_id=f"rightsize_{edge_node}_{resource_type}_{datetime.now().timestamp()}",
+            edge_node=edge_node,
+            current_setup={
+                "instance_type": "current",
+                "cost": current_cost,
+                "resource_type": resource_type,
+                "utilization": utilization,
+            },
+            recommended_setup={
+                "instance_type": "rightsized",
+                "cost": rightsized_cost,
+                "resource_type": resource_type,
+                "target_utilization": utilization_factor,
+            },
+            estimated_savings=potential_savings,
+            savings_percentage=savings_percentage,
+            confidence=0.7,
+            implementation_effort="medium",
+            risk_level=risk_level,
+            reasoning=[
+                f"Current utilization is only {utilization:.1%}",
+                f"Right-sizing can save {savings_percentage:.1f}%",
+                "Recommend gradual capacity reduction with monitoring",
+            ],
+        )
+
+    def _assess_interruption_risk(self, edge_node: str, resource_type: str) -> float:
+        """Assess interruption risk for spot instances."""
+        # Simplified risk assessment
+        # In production, this would analyze historical interruption data
+
+        base_risk = 0.15  # 15% base interruption risk
+
+        # Adjust based on resource type
+        if resource_type == "gpu":
+            base_risk *= 1.5  # GPUs have higher interruption risk
+        elif resource_type == "memory":
+            base_risk *= 0.8  # Memory instances more stable
+
+        # Adjust based on time patterns
+        # Assume we have access to usage patterns
+        peak_usage = self._is_peak_usage_time()
+        if peak_usage:
+            base_risk *= 1.3
+
+        return min(base_risk, 0.5)  # Cap at 50%
+
+    def _calculate_usage_consistency(self, edge_node: str, resource_type: str) -> float:
+        """Calculate usage consistency for reserved instance evaluation."""
+        # Analyze usage patterns over time
+        node_metrics = [
+            m
+            for m in self.cost_metrics
+            if m.edge_node == edge_node and m.resource_type == resource_type
+        ]
+
+        if len(node_metrics) < 7:  # Need at least a week of data
+            return 0.0
+
+        # Calculate daily usage
+        daily_usage = defaultdict(float)
+        for metric in node_metrics:
+            day = metric.timestamp.date()
+            daily_usage[day] += metric.usage_hours
+
+        usage_values = list(daily_usage.values())
+
+        if not usage_values:
+            return 0.0
+
+        # Calculate coefficient of variation
+        mean_usage = np.mean(usage_values)
+        std_usage = np.std(usage_values)
+
+        if mean_usage == 0:
+            return 0.0
+
+        cv = std_usage / mean_usage
+        consistency = max(0, 1 - cv)  # Lower CV = higher consistency
+
+        return consistency
+
+    def _get_resource_utilization(
+        self, edge_node: str, resource_type: str
+    ) -> Optional[float]:
+        """Get average resource utilization."""
+        # This would integrate with monitoring data
+        # For now, return simulated utilization
+
+        import random
+
+        random.seed(hash(f"{edge_node}_{resource_type}"))
+        return random.uniform(0.3, 0.9)
+
+    def _is_peak_usage_time(self) -> bool:
+        """Check if current time is peak usage."""
+        hour = datetime.now().hour
+        # Assume peak hours are 9 AM to 6 PM
+        return 9 <= hour <= 18
+
+    async def _analyze_spot_opportunity(
+        self, edge_node: str, current_costs: Dict[str, Any]
+    ) -> Optional[SpotInstanceRecommendation]:
+        """Analyze spot instance opportunity for a node."""
+        total_on_demand = sum(
+            costs.get("on_demand", 0)
+            for costs in current_costs["cost_breakdown"].values()
+        )
+
+        if total_on_demand == 0:
+            return None
+
+        # Calculate potential spot savings
+        spot_cost = total_on_demand * 0.3  # 70% savings
+        potential_savings = total_on_demand - spot_cost
+
+        # Assess interruption risk
+        interruption_risk = self._assess_interruption_risk(edge_node, "mixed")
+
+        # Determine strategy
+        if interruption_risk < 0.1:
+            strategy = "full_spot"
+        elif interruption_risk < 0.3:
+            strategy = "mixed_spot_on_demand"
+        else:
+            strategy = "diversified_spot"
+
+        # Create backup plan
+        backup_plan = {
+            "strategy": "auto_fallback",
+            "fallback_instances": "on_demand",
+            "max_interruptions_per_day": 2,
+            "auto_restart": True,
+        }
+
+        return SpotInstanceRecommendation(
+            edge_node=edge_node,
+            current_on_demand_cost=total_on_demand,
+            spot_cost=spot_cost,
+            potential_savings=potential_savings,
+            interruption_risk=interruption_risk,
+            recommended_strategy=strategy,
+            backup_plan=backup_plan,
+        )
+
+    def _analyze_provider_usage(
+        self, provider: CloudProvider
+    ) -> Dict[str, Dict[str, Any]]:
+        """Analyze usage patterns for a provider."""
+        provider_metrics = [m for m in self.cost_metrics if m.provider == provider]
+
+        usage_patterns = defaultdict(
+            lambda: {"total_hours": 0, "consistency": 0, "monthly_cost": 0}
+        )
+
+        for metric in provider_metrics:
+            patterns = usage_patterns[metric.resource_type]
+            patterns["total_hours"] += metric.usage_hours
+            patterns["monthly_cost"] += metric.total_cost
+
+        return dict(usage_patterns)
+
+    async def _analyze_reservation_opportunity(
+        self, provider: CloudProvider, resource_type: str, usage: Dict[str, Any]
+    ) -> Optional[ReservationRecommendation]:
+        """Analyze reservation opportunity for a resource."""
+        monthly_hours = usage["total_hours"]
+        monthly_cost = usage["monthly_cost"]
+
+        # Need significant usage to justify reservation
+        if monthly_hours < 500:  # Less than ~70% of month
+            return None
+
+        # Calculate reservation costs
+        hourly_rate = monthly_cost / monthly_hours if monthly_hours > 0 else 0
+
+        # 1-year reservation (30% savings)
+        reserved_hourly = hourly_rate * 0.7
+        upfront_cost = hourly_rate * monthly_hours * 0.3  # 30% upfront
+        monthly_reserved_cost = reserved_hourly * monthly_hours
+
+        total_savings = (monthly_cost - monthly_reserved_cost) * 12
+        breakeven_months = (
+            upfront_cost / (monthly_cost - monthly_reserved_cost)
+            if monthly_cost > monthly_reserved_cost
+            else 12
+        )
+
+        if total_savings <= 0 or breakeven_months > 12:
+            return None
+
+        return ReservationRecommendation(
+            resource_type=resource_type,
+            provider=provider,
+            commitment_length=12,
+            upfront_cost=upfront_cost,
+            monthly_cost=monthly_reserved_cost,
+            on_demand_equivalent=monthly_cost,
+            total_savings=total_savings,
+            breakeven_months=int(breakeven_months),
+            utilization_requirement=0.7,
+        )
+
+    def _get_risk_multiplier(self, risk_level: str) -> float:
+        """Get risk multiplier for ROI calculations."""
+        multipliers = {"low": 0.95, "medium": 0.85, "high": 0.7}
+        return multipliers.get(risk_level, 0.85)
+
+    def _get_roi_recommendation(
+        self, roi_percentage: float, payback_months: float
+    ) -> str:
+        """Get ROI-based recommendation."""
+        if roi_percentage > 100 and payback_months < 6:
+            return "Strongly Recommended"
+        elif roi_percentage > 50 and payback_months < 12:
+            return "Recommended"
+        elif roi_percentage > 20 and payback_months < 18:
+            return "Consider"
+        else:
+            return "Not Recommended"
+
+    def _calculate_current_monthly_spend(self) -> float:
+        """Calculate current monthly spend."""
+        # Get last 30 days of metrics
+        cutoff = datetime.now() - timedelta(days=30)
+        recent_metrics = [m for m in self.cost_metrics if m.timestamp > cutoff]
+
+        return sum(m.total_cost for m in recent_metrics)