diff --git a/cmd/kubelet/app/server.go b/cmd/kubelet/app/server.go
index e9c1af9470b..077ef4d41ff 100644
--- a/cmd/kubelet/app/server.go
+++ b/cmd/kubelet/app/server.go
@@ -801,7 +801,7 @@ func run(ctx context.Context, s *options.KubeletServer, kubeDeps *kubelet.Depend
 				NodeAllocatableConfig: cm.NodeAllocatableConfig{
 					KubeReservedCgroupName:   s.KubeReservedCgroup,
 					SystemReservedCgroupName: s.SystemReservedCgroup,
-					EnforceNodeAllocatable:   sets.NewString(s.EnforceNodeAllocatable...),
+					EnforceNodeAllocatable:   sets.New(s.EnforceNodeAllocatable...),
 					KubeReserved:             kubeReserved,
 					SystemReserved:           systemReserved,
 					ReservedSystemCPUs:       reservedSystemCPUs,
diff --git a/pkg/kubelet/cm/cgroup_manager_linux.go b/pkg/kubelet/cm/cgroup_manager_linux.go
index f54eaa2979f..1cd5f662429 100644
--- a/pkg/kubelet/cm/cgroup_manager_linux.go
+++ b/pkg/kubelet/cm/cgroup_manager_linux.go
@@ -246,7 +246,7 @@ func (m *cgroupManagerImpl) Validate(name CgroupName) error {
 		}
 		difference := neededControllers.Difference(enabledControllers)
 		if difference.Len() > 0 {
-			return fmt.Errorf("cgroup %q has some missing controllers: %v", name, strings.Join(difference.List(), ", "))
+			return fmt.Errorf("cgroup %q has some missing controllers: %v", name, strings.Join(sets.List(difference), ", "))
 		}
 		return nil // valid V2 cgroup
 	}
@@ -260,7 +260,7 @@ func (m *cgroupManagerImpl) Validate(name CgroupName) error {
 	// scoped to the set control groups it understands.  this is being discussed
 	// in https://github.com/opencontainers/runc/issues/1440
 	// once resolved, we can remove this code.
-	allowlistControllers := sets.NewString("cpu", "cpuacct", "cpuset", "memory", "systemd", "pids")
+	allowlistControllers := sets.New[string]("cpu", "cpuacct", "cpuset", "memory", "systemd", "pids")
 
 	if _, ok := m.subsystems.MountPoints["hugetlb"]; ok {
 		allowlistControllers.Insert("hugetlb")
@@ -322,24 +322,24 @@ func getCPUWeight(cpuShares *uint64) uint64 {
 }
 
 // readUnifiedControllers reads the controllers available at the specified cgroup
-func readUnifiedControllers(path string) (sets.String, error) {
+func readUnifiedControllers(path string) (sets.Set[string], error) {
 	controllersFileContent, err := os.ReadFile(filepath.Join(path, "cgroup.controllers"))
 	if err != nil {
 		return nil, err
 	}
 	controllers := strings.Fields(string(controllersFileContent))
-	return sets.NewString(controllers...), nil
+	return sets.New(controllers...), nil
 }
 
 var (
 	availableRootControllersOnce sync.Once
-	availableRootControllers     sets.String
+	availableRootControllers     sets.Set[string]
 )
 
 // getSupportedUnifiedControllers returns a set of supported controllers when running on cgroup v2
-func getSupportedUnifiedControllers() sets.String {
+func getSupportedUnifiedControllers() sets.Set[string] {
 	// This is the set of controllers used by the Kubelet
-	supportedControllers := sets.NewString("cpu", "cpuset", "memory", "hugetlb", "pids")
+	supportedControllers := sets.New("cpu", "cpuset", "memory", "hugetlb", "pids")
 	// Memoize the set of controllers that are present in the root cgroup
 	availableRootControllersOnce.Do(func() {
 		var err error
@@ -407,7 +407,7 @@ func (m *cgroupManagerImpl) maybeSetHugetlb(resourceConfig *ResourceConfig, reso
 	}
 
 	// For each page size enumerated, set that value.
-	pageSizes := sets.NewString()
+	pageSizes := sets.New[string]()
 	for pageSize, limit := range resourceConfig.HugePageLimit {
 		sizeString, err := v1helper.HugePageUnitSizeFromByteSize(pageSize)
 		if err != nil {
@@ -485,7 +485,7 @@ func (m *cgroupManagerImpl) Pids(name CgroupName) []int {
 	cgroupFsName := m.Name(name)
 
 	// Get a list of processes that we need to kill
-	pidsToKill := sets.NewInt()
+	pidsToKill := sets.New[int]()
 	var pids []int
 	for _, val := range m.subsystems.MountPoints {
 		dir := path.Join(val, cgroupFsName)
@@ -526,7 +526,7 @@ func (m *cgroupManagerImpl) Pids(name CgroupName) []int {
 			klog.V(4).InfoS("Cgroup manager encountered error scanning pids for directory", "path", dir, "err", err)
 		}
 	}
-	return pidsToKill.List()
+	return sets.List(pidsToKill)
 }
 
 // ReduceCPULimits reduces the cgroup's cpu shares to the lowest possible value
diff --git a/pkg/kubelet/cm/container_manager.go b/pkg/kubelet/cm/container_manager.go
index 527b66bf50e..1f8d43f8954 100644
--- a/pkg/kubelet/cm/container_manager.go
+++ b/pkg/kubelet/cm/container_manager.go
@@ -164,7 +164,7 @@ type NodeAllocatableConfig struct {
 	KubeReservedCgroupName   string
 	SystemReservedCgroupName string
 	ReservedSystemCPUs       cpuset.CPUSet
-	EnforceNodeAllocatable   sets.String
+	EnforceNodeAllocatable   sets.Set[string]
 	KubeReserved             v1.ResourceList
 	SystemReserved           v1.ResourceList
 	HardEvictionThresholds   []evictionapi.Threshold
diff --git a/pkg/kubelet/cm/container_manager_linux.go b/pkg/kubelet/cm/container_manager_linux.go
index 18001ed61f6..7dbfd4cc664 100644
--- a/pkg/kubelet/cm/container_manager_linux.go
+++ b/pkg/kubelet/cm/container_manager_linux.go
@@ -161,7 +161,7 @@ func validateSystemRequirements(mountUtil mount.Interface) (features, error) {
 		return f, nil
 	}
 
-	expectedCgroups := sets.NewString("cpu", "cpuacct", "cpuset", "memory")
+	expectedCgroups := sets.New("cpu", "cpuacct", "cpuset", "memory")
 	for _, mountPoint := range mountPoints {
 		if mountPoint.Type == cgroupMountType {
 			for _, opt := range mountPoint.Opts {
@@ -176,7 +176,7 @@ func validateSystemRequirements(mountUtil mount.Interface) (features, error) {
 	}
 
 	if expectedCgroups.Len() > 0 {
-		return f, fmt.Errorf("%s - Following Cgroup subsystem not mounted: %v", localErr, expectedCgroups.List())
+		return f, fmt.Errorf("%s - Following Cgroup subsystem not mounted: %v", localErr, sets.List(expectedCgroups))
 	}
 
 	// Check if cpu quota is available.
diff --git a/pkg/kubelet/cm/cpumanager/policy_options.go b/pkg/kubelet/cm/cpumanager/policy_options.go
index 2e275254c33..368fc63624f 100644
--- a/pkg/kubelet/cm/cpumanager/policy_options.go
+++ b/pkg/kubelet/cm/cpumanager/policy_options.go
@@ -35,14 +35,14 @@ const (
 )
 
 var (
-	alphaOptions = sets.NewString(
+	alphaOptions = sets.New[string](
 		DistributeCPUsAcrossNUMAOption,
 		AlignBySocketOption,
 	)
-	betaOptions = sets.NewString(
+	betaOptions = sets.New[string](
 		FullPCPUsOnlyOption,
 	)
-	stableOptions = sets.NewString()
+	stableOptions = sets.New[string]()
 )
 
 // CheckPolicyOptionAvailable verifies if the given option can be used depending on the Feature Gate Settings.
diff --git a/pkg/kubelet/cm/devicemanager/checkpoint/checkpoint.go b/pkg/kubelet/cm/devicemanager/checkpoint/checkpoint.go
index a8cb24be6e5..46d0aaa2a25 100644
--- a/pkg/kubelet/cm/devicemanager/checkpoint/checkpoint.go
+++ b/pkg/kubelet/cm/devicemanager/checkpoint/checkpoint.go
@@ -62,9 +62,9 @@ func NewDevicesPerNUMA() DevicesPerNUMA {
 }
 
 // Devices is a function that returns all device ids for all NUMA nodes
-// and represent it as sets.String
-func (dev DevicesPerNUMA) Devices() sets.String {
-	result := sets.NewString()
+// and represent it as sets.Set[string]
+func (dev DevicesPerNUMA) Devices() sets.Set[string] {
+	result := sets.New[string]()
 
 	for _, devs := range dev {
 		result.Insert(devs...)
diff --git a/pkg/kubelet/cm/devicemanager/manager.go b/pkg/kubelet/cm/devicemanager/manager.go
index d780ee801bd..3808ccfa0d3 100644
--- a/pkg/kubelet/cm/devicemanager/manager.go
+++ b/pkg/kubelet/cm/devicemanager/manager.go
@@ -74,13 +74,13 @@ type ManagerImpl struct {
 	allDevices ResourceDeviceInstances
 
 	// healthyDevices contains all of the registered healthy resourceNames and their exported device IDs.
-	healthyDevices map[string]sets.String
+	healthyDevices map[string]sets.Set[string]
 
 	// unhealthyDevices contains all of the unhealthy devices and their exported device IDs.
-	unhealthyDevices map[string]sets.String
+	unhealthyDevices map[string]sets.Set[string]
 
 	// allocatedDevices contains allocated deviceIds, keyed by resourceName.
-	allocatedDevices map[string]sets.String
+	allocatedDevices map[string]sets.Set[string]
 
 	// podDevices contains pod to allocated device mapping.
 	podDevices        *podDevices
@@ -106,7 +106,7 @@ type ManagerImpl struct {
 	// containerRunningSet identifies which container among those present in `containerMap`
 	// was reported running by the container runtime when `containerMap` was computed.
 	// Used to detect pods running across a restart
-	containerRunningSet sets.String
+	containerRunningSet sets.Set[string]
 }
 
 type endpointInfo struct {
@@ -117,7 +117,7 @@ type endpointInfo struct {
 type sourcesReadyStub struct{}
 
 // PodReusableDevices is a map by pod name of devices to reuse.
-type PodReusableDevices map[string]map[string]sets.String
+type PodReusableDevices map[string]map[string]sets.Set[string]
 
 func (s *sourcesReadyStub) AddSource(source string) {}
 func (s *sourcesReadyStub) AllReady() bool          { return true }
@@ -143,9 +143,9 @@ func newManagerImpl(socketPath string, topology []cadvisorapi.Node, topologyAffi
 		endpoints: make(map[string]endpointInfo),
 
 		allDevices:            NewResourceDeviceInstances(),
-		healthyDevices:        make(map[string]sets.String),
-		unhealthyDevices:      make(map[string]sets.String),
-		allocatedDevices:      make(map[string]sets.String),
+		healthyDevices:        make(map[string]sets.Set[string]),
+		unhealthyDevices:      make(map[string]sets.Set[string]),
+		allocatedDevices:      make(map[string]sets.Set[string]),
 		podDevices:            newPodDevices(),
 		numaNodes:             numaNodes,
 		topologyAffinityStore: topologyAffinityStore,
@@ -259,8 +259,8 @@ func (m *ManagerImpl) PluginListAndWatchReceiver(resourceName string, resp *plug
 func (m *ManagerImpl) genericDeviceUpdateCallback(resourceName string, devices []pluginapi.Device) {
 	healthyCount := 0
 	m.mutex.Lock()
-	m.healthyDevices[resourceName] = sets.NewString()
-	m.unhealthyDevices[resourceName] = sets.NewString()
+	m.healthyDevices[resourceName] = sets.New[string]()
+	m.unhealthyDevices[resourceName] = sets.New[string]()
 	m.allDevices[resourceName] = make(map[string]pluginapi.Device)
 	for _, dev := range devices {
 		m.allDevices[resourceName][dev.ID] = dev
@@ -291,7 +291,7 @@ func (m *ManagerImpl) checkpointFile() string {
 // Start starts the Device Plugin Manager and start initialization of
 // podDevices and allocatedDevices information from checkpointed state and
 // starts device plugin registration service.
-func (m *ManagerImpl) Start(activePods ActivePodsFunc, sourcesReady config.SourcesReady, initialContainers containermap.ContainerMap, initialContainerRunningSet sets.String) error {
+func (m *ManagerImpl) Start(activePods ActivePodsFunc, sourcesReady config.SourcesReady, initialContainers containermap.ContainerMap, initialContainerRunningSet sets.Set[string]) error {
 	klog.V(2).InfoS("Starting Device Plugin manager")
 
 	m.activePods = activePods
@@ -323,7 +323,7 @@ func (m *ManagerImpl) Allocate(pod *v1.Pod, container *v1.Container) error {
 	m.setPodPendingAdmission(pod)
 
 	if _, ok := m.devicesToReuse[string(pod.UID)]; !ok {
-		m.devicesToReuse[string(pod.UID)] = make(map[string]sets.String)
+		m.devicesToReuse[string(pod.UID)] = make(map[string]sets.Set[string])
 	}
 	// If pod entries to m.devicesToReuse other than the current pod exist, delete them.
 	for podUID := range m.devicesToReuse {
@@ -365,13 +365,13 @@ func (m *ManagerImpl) UpdatePluginResources(node *schedulerframework.NodeInfo, a
 
 func (m *ManagerImpl) markResourceUnhealthy(resourceName string) {
 	klog.V(2).InfoS("Mark all resources Unhealthy for resource", "resourceName", resourceName)
-	healthyDevices := sets.NewString()
+	healthyDevices := sets.New[string]()
 	if _, ok := m.healthyDevices[resourceName]; ok {
 		healthyDevices = m.healthyDevices[resourceName]
-		m.healthyDevices[resourceName] = sets.NewString()
+		m.healthyDevices[resourceName] = sets.New[string]()
 	}
 	if _, ok := m.unhealthyDevices[resourceName]; !ok {
-		m.unhealthyDevices[resourceName] = sets.NewString()
+		m.unhealthyDevices[resourceName] = sets.New[string]()
 	}
 	m.unhealthyDevices[resourceName] = m.unhealthyDevices[resourceName].Union(healthyDevices)
 }
@@ -392,7 +392,7 @@ func (m *ManagerImpl) GetCapacity() (v1.ResourceList, v1.ResourceList, []string)
 	needsUpdateCheckpoint := false
 	var capacity = v1.ResourceList{}
 	var allocatable = v1.ResourceList{}
-	deletedResources := sets.NewString()
+	deletedResources := sets.New[string]()
 	m.mutex.Lock()
 	for resourceName, devices := range m.healthyDevices {
 		eI, ok := m.endpoints[resourceName]
@@ -492,8 +492,8 @@ func (m *ManagerImpl) readCheckpoint() error {
 	for resource := range registeredDevs {
 		// During start up, creates empty healthyDevices list so that the resource capacity
 		// will stay zero till the corresponding device plugin re-registers.
-		m.healthyDevices[resource] = sets.NewString()
-		m.unhealthyDevices[resource] = sets.NewString()
+		m.healthyDevices[resource] = sets.New[string]()
+		m.unhealthyDevices[resource] = sets.New[string]()
 		m.endpoints[resource] = endpointInfo{e: newStoppedEndpointImpl(resource), opts: nil}
 	}
 	return nil
@@ -536,15 +536,15 @@ func (m *ManagerImpl) UpdateAllocatedDevices() {
 	if len(podsToBeRemoved) <= 0 {
 		return
 	}
-	klog.V(3).InfoS("Pods to be removed", "podUIDs", podsToBeRemoved.List())
-	m.podDevices.delete(podsToBeRemoved.List())
+	klog.V(3).InfoS("Pods to be removed", "podUIDs", sets.List(podsToBeRemoved))
+	m.podDevices.delete(sets.List(podsToBeRemoved))
 	// Regenerated allocatedDevices after we update pod allocation information.
 	m.allocatedDevices = m.podDevices.devices()
 }
 
 // Returns list of device Ids we need to allocate with Allocate rpc call.
 // Returns empty list in case we don't need to issue the Allocate rpc call.
-func (m *ManagerImpl) devicesToAllocate(podUID, contName, resource string, required int, reusableDevices sets.String) (sets.String, error) {
+func (m *ManagerImpl) devicesToAllocate(podUID, contName, resource string, required int, reusableDevices sets.Set[string]) (sets.Set[string], error) {
 	m.mutex.Lock()
 	defer m.mutex.Unlock()
 	needed := required
@@ -552,7 +552,7 @@ func (m *ManagerImpl) devicesToAllocate(podUID, contName, resource string, requi
 	// This can happen if a container restarts for example.
 	devices := m.podDevices.containerDevices(podUID, contName, resource)
 	if devices != nil {
-		klog.V(3).InfoS("Found pre-allocated devices for resource on pod", "resourceName", resource, "containerName", contName, "podUID", string(podUID), "devices", devices.List())
+		klog.V(3).InfoS("Found pre-allocated devices for resource on pod", "resourceName", resource, "containerName", contName, "podUID", string(podUID), "devices", sets.List(devices))
 		needed = needed - devices.Len()
 		// A pod's resource is not expected to change once admitted by the API server,
 		// so just fail loudly here. We can revisit this part if this no longer holds.
@@ -610,11 +610,11 @@ func (m *ManagerImpl) devicesToAllocate(podUID, contName, resource string, requi
 
 	// Declare the list of allocated devices.
 	// This will be populated and returned below.
-	allocated := sets.NewString()
+	allocated := sets.New[string]()
 
 	// Create a closure to help with device allocation
 	// Returns 'true' once no more devices need to be allocated.
-	allocateRemainingFrom := func(devices sets.String) bool {
+	allocateRemainingFrom := func(devices sets.Set[string]) bool {
 		for device := range devices.Difference(allocated) {
 			m.allocatedDevices[resource].Insert(device)
 			allocated.Insert(device)
@@ -628,7 +628,7 @@ func (m *ManagerImpl) devicesToAllocate(podUID, contName, resource string, requi
 
 	// Needs to allocate additional devices.
 	if m.allocatedDevices[resource] == nil {
-		m.allocatedDevices[resource] = sets.NewString()
+		m.allocatedDevices[resource] = sets.New[string]()
 	}
 
 	// Allocates from reusableDevices list first.
@@ -697,22 +697,22 @@ func (m *ManagerImpl) devicesToAllocate(podUID, contName, resource string, requi
 	return nil, fmt.Errorf("unexpectedly allocated less resources than required. Requested: %d, Got: %d", required, required-needed)
 }
 
-func (m *ManagerImpl) filterByAffinity(podUID, contName, resource string, available sets.String) (sets.String, sets.String, sets.String) {
+func (m *ManagerImpl) filterByAffinity(podUID, contName, resource string, available sets.Set[string]) (sets.Set[string], sets.Set[string], sets.Set[string]) {
 	// If alignment information is not available, just pass the available list back.
 	hint := m.topologyAffinityStore.GetAffinity(podUID, contName)
 	if !m.deviceHasTopologyAlignment(resource) || hint.NUMANodeAffinity == nil {
-		return sets.NewString(), sets.NewString(), available
+		return sets.New[string](), sets.New[string](), available
 	}
 
 	// Build a map of NUMA Nodes to the devices associated with them. A
 	// device may be associated to multiple NUMA nodes at the same time. If an
 	// available device does not have any NUMA Nodes associated with it, add it
 	// to a list of NUMA Nodes for the fake NUMANode -1.
-	perNodeDevices := make(map[int]sets.String)
+	perNodeDevices := make(map[int]sets.Set[string])
 	for d := range available {
 		if m.allDevices[resource][d].Topology == nil || len(m.allDevices[resource][d].Topology.Nodes) == 0 {
 			if _, ok := perNodeDevices[nodeWithoutTopology]; !ok {
-				perNodeDevices[nodeWithoutTopology] = sets.NewString()
+				perNodeDevices[nodeWithoutTopology] = sets.New[string]()
 			}
 			perNodeDevices[nodeWithoutTopology].Insert(d)
 			continue
@@ -720,7 +720,7 @@ func (m *ManagerImpl) filterByAffinity(podUID, contName, resource string, availa
 
 		for _, node := range m.allDevices[resource][d].Topology.Nodes {
 			if _, ok := perNodeDevices[int(node.ID)]; !ok {
-				perNodeDevices[int(node.ID)] = sets.NewString()
+				perNodeDevices[int(node.ID)] = sets.New[string]()
 			}
 			perNodeDevices[int(node.ID)].Insert(d)
 		}
@@ -791,14 +791,14 @@ func (m *ManagerImpl) filterByAffinity(podUID, contName, resource string, availa
 	}
 
 	// Return all three lists containing the full set of devices across them.
-	return sets.NewString(fromAffinity...), sets.NewString(notFromAffinity...), sets.NewString(withoutTopology...)
+	return sets.New[string](fromAffinity...), sets.New[string](notFromAffinity...), sets.New[string](withoutTopology...)
 }
 
 // allocateContainerResources attempts to allocate all of required device
 // plugin resources for the input container, issues an Allocate rpc request
 // for each new device resource requirement, processes their AllocateResponses,
 // and updates the cached containerDevices on success.
-func (m *ManagerImpl) allocateContainerResources(pod *v1.Pod, container *v1.Container, devicesToReuse map[string]sets.String) error {
+func (m *ManagerImpl) allocateContainerResources(pod *v1.Pod, container *v1.Container, devicesToReuse map[string]sets.Set[string]) error {
 	podUID := string(pod.UID)
 	contName := container.Name
 	allocatedDevicesUpdated := false
@@ -981,7 +981,7 @@ func (m *ManagerImpl) callPreStartContainerIfNeeded(podUID, contName, resource s
 
 // callGetPreferredAllocationIfAvailable issues GetPreferredAllocation grpc
 // call for device plugin resource with GetPreferredAllocationAvailable option set.
-func (m *ManagerImpl) callGetPreferredAllocationIfAvailable(podUID, contName, resource string, available, mustInclude sets.String, size int) (sets.String, error) {
+func (m *ManagerImpl) callGetPreferredAllocationIfAvailable(podUID, contName, resource string, available, mustInclude sets.Set[string], size int) (sets.Set[string], error) {
 	eI, ok := m.endpoints[resource]
 	if !ok {
 		return nil, fmt.Errorf("endpoint not found in cache for a registered resource: %s", resource)
@@ -1000,9 +1000,9 @@ func (m *ManagerImpl) callGetPreferredAllocationIfAvailable(podUID, contName, re
 		return nil, fmt.Errorf("device plugin GetPreferredAllocation rpc failed with err: %v", err)
 	}
 	if resp != nil && len(resp.ContainerResponses) > 0 {
-		return sets.NewString(resp.ContainerResponses[0].DeviceIDs...), nil
+		return sets.New[string](resp.ContainerResponses[0].DeviceIDs...), nil
 	}
-	return sets.NewString(), nil
+	return sets.New[string](), nil
 }
 
 // sanitizeNodeAllocatable scans through allocatedDevices in the device manager
diff --git a/pkg/kubelet/cm/devicemanager/manager_test.go b/pkg/kubelet/cm/devicemanager/manager_test.go
index 4a23e37858c..7eb9c53356e 100644
--- a/pkg/kubelet/cm/devicemanager/manager_test.go
+++ b/pkg/kubelet/cm/devicemanager/manager_test.go
@@ -287,7 +287,7 @@ func setupDeviceManager(t *testing.T, devs []*pluginapi.Device, callback monitor
 
 	// test steady state, initialization where sourcesReady, containerMap and containerRunningSet
 	// are relevant will be tested with a different flow
-	err = w.Start(activePods, &sourcesReadyStub{}, containermap.NewContainerMap(), sets.NewString())
+	err = w.Start(activePods, &sourcesReadyStub{}, containermap.NewContainerMap(), sets.New[string]())
 	require.NoError(t, err)
 
 	return w, updateChan
@@ -312,6 +312,7 @@ func setupPluginManager(t *testing.T, pluginSocketName string, m Manager) plugin
 }
 
 func runPluginManager(pluginManager pluginmanager.PluginManager) {
+	// FIXME: Replace sets.String with sets.Set[string]
 	sourcesReady := config.NewSourcesReady(func(_ sets.String) bool { return true })
 	go pluginManager.Run(sourcesReady, wait.NeverStop)
 }
@@ -459,8 +460,8 @@ func TestUpdateCapacityAllocatable(t *testing.T) {
 	// properly rejected instead of being incorrectly started.
 	err = testManager.writeCheckpoint()
 	as.Nil(err)
-	testManager.healthyDevices = make(map[string]sets.String)
-	testManager.unhealthyDevices = make(map[string]sets.String)
+	testManager.healthyDevices = make(map[string]sets.Set[string])
+	testManager.unhealthyDevices = make(map[string]sets.Set[string])
 	err = testManager.readCheckpoint()
 	as.Nil(err)
 	as.Equal(1, len(testManager.endpoints))
@@ -673,9 +674,9 @@ func TestCheckpoint(t *testing.T) {
 	as.Nil(err)
 	testManager := &ManagerImpl{
 		endpoints:         make(map[string]endpointInfo),
-		healthyDevices:    make(map[string]sets.String),
-		unhealthyDevices:  make(map[string]sets.String),
-		allocatedDevices:  make(map[string]sets.String),
+		healthyDevices:    make(map[string]sets.Set[string]),
+		unhealthyDevices:  make(map[string]sets.Set[string]),
+		allocatedDevices:  make(map[string]sets.Set[string]),
 		podDevices:        newPodDevices(),
 		checkpointManager: ckm,
 	}
@@ -718,16 +719,16 @@ func TestCheckpoint(t *testing.T) {
 		),
 	)
 
-	testManager.healthyDevices[resourceName1] = sets.NewString()
+	testManager.healthyDevices[resourceName1] = sets.New[string]()
 	testManager.healthyDevices[resourceName1].Insert("dev1")
 	testManager.healthyDevices[resourceName1].Insert("dev2")
 	testManager.healthyDevices[resourceName1].Insert("dev3")
 	testManager.healthyDevices[resourceName1].Insert("dev4")
 	testManager.healthyDevices[resourceName1].Insert("dev5")
-	testManager.healthyDevices[resourceName2] = sets.NewString()
+	testManager.healthyDevices[resourceName2] = sets.New[string]()
 	testManager.healthyDevices[resourceName2].Insert("dev1")
 	testManager.healthyDevices[resourceName2].Insert("dev2")
-	testManager.healthyDevices[resourceName3] = sets.NewString()
+	testManager.healthyDevices[resourceName3] = sets.New[string]()
 	testManager.healthyDevices[resourceName3].Insert("dev5")
 
 	expectedPodDevices := testManager.podDevices
@@ -827,9 +828,9 @@ func getTestManager(tmpDir string, activePods ActivePodsFunc, testRes []TestReso
 		return nil, err
 	}
 	m := &ManagerImpl{
-		healthyDevices:        make(map[string]sets.String),
-		unhealthyDevices:      make(map[string]sets.String),
-		allocatedDevices:      make(map[string]sets.String),
+		healthyDevices:        make(map[string]sets.Set[string]),
+		unhealthyDevices:      make(map[string]sets.Set[string]),
+		allocatedDevices:      make(map[string]sets.Set[string]),
 		endpoints:             make(map[string]endpointInfo),
 		podDevices:            newPodDevices(),
 		devicesToReuse:        make(PodReusableDevices),
@@ -846,7 +847,7 @@ func getTestManager(tmpDir string, activePods ActivePodsFunc, testRes []TestReso
 	}
 
 	for _, res := range testRes {
-		testManager.healthyDevices[res.resourceName] = sets.NewString(res.devs.Devices().UnsortedList()...)
+		testManager.healthyDevices[res.resourceName] = sets.New[string](res.devs.Devices().UnsortedList()...)
 		if res.resourceName == "domain1.com/resource1" {
 			testManager.endpoints[res.resourceName] = endpointInfo{
 				e:    &MockEndpoint{allocateFunc: allocateStubFunc()},
@@ -953,22 +954,22 @@ func TestFilterByAffinity(t *testing.T) {
 	}
 
 	testCases := []struct {
-		available               sets.String
-		fromAffinityExpected    sets.String
-		notFromAffinityExpected sets.String
-		withoutTopologyExpected sets.String
+		available               sets.Set[string]
+		fromAffinityExpected    sets.Set[string]
+		notFromAffinityExpected sets.Set[string]
+		withoutTopologyExpected sets.Set[string]
 	}{
 		{
-			available:               sets.NewString("dev1", "dev2"),
-			fromAffinityExpected:    sets.NewString("dev2"),
-			notFromAffinityExpected: sets.NewString("dev1"),
-			withoutTopologyExpected: sets.NewString(),
+			available:               sets.New[string]("dev1", "dev2"),
+			fromAffinityExpected:    sets.New[string]("dev2"),
+			notFromAffinityExpected: sets.New[string]("dev1"),
+			withoutTopologyExpected: sets.New[string](),
 		},
 		{
-			available:               sets.NewString("dev1", "dev2", "dev3", "dev4"),
-			fromAffinityExpected:    sets.NewString("dev2", "dev3", "dev4"),
-			notFromAffinityExpected: sets.NewString("dev1"),
-			withoutTopologyExpected: sets.NewString(),
+			available:               sets.New[string]("dev1", "dev2", "dev3", "dev4"),
+			fromAffinityExpected:    sets.New[string]("dev2", "dev3", "dev4"),
+			notFromAffinityExpected: sets.New[string]("dev1"),
+			withoutTopologyExpected: sets.New[string](),
 		},
 	}
 
@@ -1087,9 +1088,9 @@ func TestPodContainerDeviceToAllocate(t *testing.T) {
 
 	testManager := &ManagerImpl{
 		endpoints:        make(map[string]endpointInfo),
-		healthyDevices:   make(map[string]sets.String),
-		unhealthyDevices: make(map[string]sets.String),
-		allocatedDevices: make(map[string]sets.String),
+		healthyDevices:   make(map[string]sets.Set[string]),
+		unhealthyDevices: make(map[string]sets.Set[string]),
+		allocatedDevices: make(map[string]sets.Set[string]),
 		podDevices:       newPodDevices(),
 		activePods:       func() []*v1.Pod { return []*v1.Pod{} },
 		sourcesReady:     &sourcesReadyStub{},
@@ -1121,8 +1122,8 @@ func TestPodContainerDeviceToAllocate(t *testing.T) {
 	// no healthy devices for resourceName1 and devices corresponding to
 	// resource2 are intentionally omitted to simulate that the resource
 	// hasn't been registered.
-	testManager.healthyDevices[resourceName1] = sets.NewString()
-	testManager.healthyDevices[resourceName3] = sets.NewString()
+	testManager.healthyDevices[resourceName1] = sets.New[string]()
+	testManager.healthyDevices[resourceName3] = sets.New[string]()
 	// dev5 is no longer in the list of healthy devices
 	testManager.healthyDevices[resourceName3].Insert("dev7")
 	testManager.healthyDevices[resourceName3].Insert("dev8")
@@ -1133,8 +1134,8 @@ func TestPodContainerDeviceToAllocate(t *testing.T) {
 		contName                 string
 		resource                 string
 		required                 int
-		reusableDevices          sets.String
-		expectedAllocatedDevices sets.String
+		reusableDevices          sets.Set[string]
+		expectedAllocatedDevices sets.Set[string]
 		expErr                   error
 	}{
 		{
@@ -1143,7 +1144,7 @@ func TestPodContainerDeviceToAllocate(t *testing.T) {
 			contName:                 "con1",
 			resource:                 resourceName1,
 			required:                 2,
-			reusableDevices:          sets.NewString(),
+			reusableDevices:          sets.New[string](),
 			expectedAllocatedDevices: nil,
 			expErr:                   fmt.Errorf("no healthy devices present; cannot allocate unhealthy devices %s", resourceName1),
 		},
@@ -1153,7 +1154,7 @@ func TestPodContainerDeviceToAllocate(t *testing.T) {
 			contName:                 "con2",
 			resource:                 resourceName2,
 			required:                 1,
-			reusableDevices:          sets.NewString(),
+			reusableDevices:          sets.New[string](),
 			expectedAllocatedDevices: nil,
 			expErr:                   fmt.Errorf("cannot allocate unregistered device %s", resourceName2),
 		},
@@ -1163,7 +1164,7 @@ func TestPodContainerDeviceToAllocate(t *testing.T) {
 			contName:                 "con3",
 			resource:                 resourceName3,
 			required:                 1,
-			reusableDevices:          sets.NewString(),
+			reusableDevices:          sets.New[string](),
 			expectedAllocatedDevices: nil,
 			expErr:                   fmt.Errorf("previously allocated devices are no longer healthy; cannot allocate unhealthy devices %s", resourceName3),
 		},
@@ -1366,8 +1367,8 @@ func TestUpdatePluginResources(t *testing.T) {
 	ckm, err := checkpointmanager.NewCheckpointManager(tmpDir)
 	as.Nil(err)
 	m := &ManagerImpl{
-		allocatedDevices:  make(map[string]sets.String),
-		healthyDevices:    make(map[string]sets.String),
+		allocatedDevices:  make(map[string]sets.Set[string]),
+		healthyDevices:    make(map[string]sets.Set[string]),
 		podDevices:        newPodDevices(),
 		checkpointManager: ckm,
 	}
@@ -1378,9 +1379,9 @@ func TestUpdatePluginResources(t *testing.T) {
 	testManager.podDevices.devs[string(pod.UID)] = make(containerDevices)
 
 	// require one of resource1 and one of resource2
-	testManager.allocatedDevices[resourceName1] = sets.NewString()
+	testManager.allocatedDevices[resourceName1] = sets.New[string]()
 	testManager.allocatedDevices[resourceName1].Insert(devID1)
-	testManager.allocatedDevices[resourceName2] = sets.NewString()
+	testManager.allocatedDevices[resourceName2] = sets.New[string]()
 	testManager.allocatedDevices[resourceName2].Insert(devID2)
 
 	cachedNode := &v1.Node{
@@ -1486,9 +1487,9 @@ func TestResetExtendedResource(t *testing.T) {
 	as.Nil(err)
 	testManager := &ManagerImpl{
 		endpoints:         make(map[string]endpointInfo),
-		healthyDevices:    make(map[string]sets.String),
-		unhealthyDevices:  make(map[string]sets.String),
-		allocatedDevices:  make(map[string]sets.String),
+		healthyDevices:    make(map[string]sets.Set[string]),
+		unhealthyDevices:  make(map[string]sets.Set[string]),
+		allocatedDevices:  make(map[string]sets.Set[string]),
 		podDevices:        newPodDevices(),
 		checkpointManager: ckm,
 	}
@@ -1502,7 +1503,7 @@ func TestResetExtendedResource(t *testing.T) {
 		),
 	)
 
-	testManager.healthyDevices[extendedResourceName] = sets.NewString()
+	testManager.healthyDevices[extendedResourceName] = sets.New[string]()
 	testManager.healthyDevices[extendedResourceName].Insert("dev1")
 	// checkpoint is present, indicating node hasn't been recreated
 	err = testManager.writeCheckpoint()
diff --git a/pkg/kubelet/cm/devicemanager/pod_devices.go b/pkg/kubelet/cm/devicemanager/pod_devices.go
index fe4eb65e405..a6b88ce4f70 100644
--- a/pkg/kubelet/cm/devicemanager/pod_devices.go
+++ b/pkg/kubelet/cm/devicemanager/pod_devices.go
@@ -52,10 +52,10 @@ func newPodDevices() *podDevices {
 	return &podDevices{devs: make(map[string]containerDevices)}
 }
 
-func (pdev *podDevices) pods() sets.String {
+func (pdev *podDevices) pods() sets.Set[string] {
 	pdev.RLock()
 	defer pdev.RUnlock()
-	ret := sets.NewString()
+	ret := sets.New[string]()
 	for k := range pdev.devs {
 		ret.Insert(k)
 	}
@@ -100,11 +100,11 @@ func (pdev *podDevices) delete(pods []string) {
 
 // Returns list of device Ids allocated to the given pod for the given resource.
 // Returns nil if we don't have cached state for the given <podUID, resource>.
-func (pdev *podDevices) podDevices(podUID, resource string) sets.String {
+func (pdev *podDevices) podDevices(podUID, resource string) sets.Set[string] {
 	pdev.RLock()
 	defer pdev.RUnlock()
 
-	ret := sets.NewString()
+	ret := sets.New[string]()
 	for contName := range pdev.devs[podUID] {
 		ret = ret.Union(pdev.containerDevices(podUID, contName, resource))
 	}
@@ -113,7 +113,7 @@ func (pdev *podDevices) podDevices(podUID, resource string) sets.String {
 
 // Returns list of device Ids allocated to the given container for the given resource.
 // Returns nil if we don't have cached state for the given <podUID, contName, resource>.
-func (pdev *podDevices) containerDevices(podUID, contName, resource string) sets.String {
+func (pdev *podDevices) containerDevices(podUID, contName, resource string) sets.Set[string] {
 	pdev.RLock()
 	defer pdev.RUnlock()
 	if _, podExists := pdev.devs[podUID]; !podExists {
@@ -130,7 +130,7 @@ func (pdev *podDevices) containerDevices(podUID, contName, resource string) sets
 }
 
 // Populates allocatedResources with the device resources allocated to the specified <podUID, contName>.
-func (pdev *podDevices) addContainerAllocatedResources(podUID, contName string, allocatedResources map[string]sets.String) {
+func (pdev *podDevices) addContainerAllocatedResources(podUID, contName string, allocatedResources map[string]sets.Set[string]) {
 	pdev.RLock()
 	defer pdev.RUnlock()
 	containers, exists := pdev.devs[podUID]
@@ -147,7 +147,7 @@ func (pdev *podDevices) addContainerAllocatedResources(podUID, contName string,
 }
 
 // Removes the device resources allocated to the specified <podUID, contName> from allocatedResources.
-func (pdev *podDevices) removeContainerAllocatedResources(podUID, contName string, allocatedResources map[string]sets.String) {
+func (pdev *podDevices) removeContainerAllocatedResources(podUID, contName string, allocatedResources map[string]sets.Set[string]) {
 	pdev.RLock()
 	defer pdev.RUnlock()
 	containers, exists := pdev.devs[podUID]
@@ -164,15 +164,15 @@ func (pdev *podDevices) removeContainerAllocatedResources(podUID, contName strin
 }
 
 // Returns all of devices allocated to the pods being tracked, keyed by resourceName.
-func (pdev *podDevices) devices() map[string]sets.String {
-	ret := make(map[string]sets.String)
+func (pdev *podDevices) devices() map[string]sets.Set[string] {
+	ret := make(map[string]sets.Set[string])
 	pdev.RLock()
 	defer pdev.RUnlock()
 	for _, containerDevices := range pdev.devs {
 		for _, resources := range containerDevices {
 			for resource, devices := range resources {
 				if _, exists := ret[resource]; !exists {
-					ret[resource] = sets.NewString()
+					ret[resource] = sets.New[string]()
 				}
 				if devices.allocResp != nil {
 					ret[resource] = ret[resource].Union(devices.deviceIds.Devices())
@@ -464,9 +464,9 @@ func (rdev ResourceDeviceInstances) Clone() ResourceDeviceInstances {
 	return clone
 }
 
-// Filter takes a condition set expressed as map[string]sets.String and returns a new
+// Filter takes a condition set expressed as map[string]sets.Set[string] and returns a new
 // ResourceDeviceInstances with only the devices matching the condition set.
-func (rdev ResourceDeviceInstances) Filter(cond map[string]sets.String) ResourceDeviceInstances {
+func (rdev ResourceDeviceInstances) Filter(cond map[string]sets.Set[string]) ResourceDeviceInstances {
 	filtered := NewResourceDeviceInstances()
 	for resourceName, filterIDs := range cond {
 		if _, exists := rdev[resourceName]; !exists {
diff --git a/pkg/kubelet/cm/devicemanager/pod_devices_test.go b/pkg/kubelet/cm/devicemanager/pod_devices_test.go
index 70d5f6e6dea..e6baf5f32db 100644
--- a/pkg/kubelet/cm/devicemanager/pod_devices_test.go
+++ b/pkg/kubelet/cm/devicemanager/pod_devices_test.go
@@ -65,7 +65,7 @@ func TestGetContainerDevices(t *testing.T) {
 
 func TestResourceDeviceInstanceFilter(t *testing.T) {
 	var expected string
-	var cond map[string]sets.String
+	var cond map[string]sets.Set[string]
 	var resp ResourceDeviceInstances
 	devs := ResourceDeviceInstances{
 		"foo": DeviceInstances{
@@ -103,40 +103,40 @@ func TestResourceDeviceInstanceFilter(t *testing.T) {
 		},
 	}
 
-	resp = devs.Filter(map[string]sets.String{})
+	resp = devs.Filter(map[string]sets.Set[string]{})
 	expected = `{}`
 	expectResourceDeviceInstances(t, resp, expected)
 
-	cond = map[string]sets.String{
-		"foo": sets.NewString("dev-foo1", "dev-foo2"),
-		"bar": sets.NewString("dev-bar1"),
+	cond = map[string]sets.Set[string]{
+		"foo": sets.New[string]("dev-foo1", "dev-foo2"),
+		"bar": sets.New[string]("dev-bar1"),
 	}
 	resp = devs.Filter(cond)
 	expected = `{"bar":{"dev-bar1":{"ID":"bar1"}},"foo":{"dev-foo1":{"ID":"foo1"},"dev-foo2":{"ID":"foo2"}}}`
 	expectResourceDeviceInstances(t, resp, expected)
 
-	cond = map[string]sets.String{
-		"foo": sets.NewString("dev-foo1", "dev-foo2", "dev-foo3"),
-		"bar": sets.NewString("dev-bar1", "dev-bar2", "dev-bar3"),
-		"baz": sets.NewString("dev-baz1", "dev-baz2", "dev-baz3"),
+	cond = map[string]sets.Set[string]{
+		"foo": sets.New[string]("dev-foo1", "dev-foo2", "dev-foo3"),
+		"bar": sets.New[string]("dev-bar1", "dev-bar2", "dev-bar3"),
+		"baz": sets.New[string]("dev-baz1", "dev-baz2", "dev-baz3"),
 	}
 	resp = devs.Filter(cond)
 	expected = `{"bar":{"dev-bar1":{"ID":"bar1"},"dev-bar2":{"ID":"bar2"},"dev-bar3":{"ID":"bar3"}},"baz":{"dev-baz1":{"ID":"baz1"},"dev-baz2":{"ID":"baz2"},"dev-baz3":{"ID":"baz3"}},"foo":{"dev-foo1":{"ID":"foo1"},"dev-foo2":{"ID":"foo2"},"dev-foo3":{"ID":"foo3"}}}`
 	expectResourceDeviceInstances(t, resp, expected)
 
-	cond = map[string]sets.String{
-		"foo": sets.NewString("dev-foo1", "dev-foo2", "dev-foo3", "dev-foo4"),
-		"bar": sets.NewString("dev-bar1", "dev-bar2", "dev-bar3", "dev-bar4"),
-		"baz": sets.NewString("dev-baz1", "dev-baz2", "dev-baz3", "dev-bar4"),
+	cond = map[string]sets.Set[string]{
+		"foo": sets.New[string]("dev-foo1", "dev-foo2", "dev-foo3", "dev-foo4"),
+		"bar": sets.New[string]("dev-bar1", "dev-bar2", "dev-bar3", "dev-bar4"),
+		"baz": sets.New[string]("dev-baz1", "dev-baz2", "dev-baz3", "dev-bar4"),
 	}
 	resp = devs.Filter(cond)
 	expected = `{"bar":{"dev-bar1":{"ID":"bar1"},"dev-bar2":{"ID":"bar2"},"dev-bar3":{"ID":"bar3"}},"baz":{"dev-baz1":{"ID":"baz1"},"dev-baz2":{"ID":"baz2"},"dev-baz3":{"ID":"baz3"}},"foo":{"dev-foo1":{"ID":"foo1"},"dev-foo2":{"ID":"foo2"},"dev-foo3":{"ID":"foo3"}}}`
 	expectResourceDeviceInstances(t, resp, expected)
 
-	cond = map[string]sets.String{
-		"foo": sets.NewString("dev-foo1", "dev-foo4", "dev-foo7"),
-		"bar": sets.NewString("dev-bar1", "dev-bar4", "dev-bar7"),
-		"baz": sets.NewString("dev-baz1", "dev-baz4", "dev-baz7"),
+	cond = map[string]sets.Set[string]{
+		"foo": sets.New[string]("dev-foo1", "dev-foo4", "dev-foo7"),
+		"bar": sets.New[string]("dev-bar1", "dev-bar4", "dev-bar7"),
+		"baz": sets.New[string]("dev-baz1", "dev-baz4", "dev-baz7"),
 	}
 	resp = devs.Filter(cond)
 	expected = `{"bar":{"dev-bar1":{"ID":"bar1"}},"baz":{"dev-baz1":{"ID":"baz1"}},"foo":{"dev-foo1":{"ID":"foo1"}}}`
diff --git a/pkg/kubelet/cm/devicemanager/topology_hints.go b/pkg/kubelet/cm/devicemanager/topology_hints.go
index d68febd2f05..12f515d3658 100644
--- a/pkg/kubelet/cm/devicemanager/topology_hints.go
+++ b/pkg/kubelet/cm/devicemanager/topology_hints.go
@@ -63,7 +63,7 @@ func (m *ManagerImpl) GetTopologyHints(pod *v1.Pod, container *v1.Container) map
 				continue
 			}
 			klog.InfoS("Regenerating TopologyHints for resource already allocated to pod", "resource", resource, "pod", klog.KObj(pod), "containerName", container.Name)
-			deviceHints[resource] = m.generateDeviceTopologyHints(resource, allocated, sets.String{}, requested)
+			deviceHints[resource] = m.generateDeviceTopologyHints(resource, allocated, sets.Set[string]{}, requested)
 			continue
 		}
 
@@ -118,7 +118,7 @@ func (m *ManagerImpl) GetPodTopologyHints(pod *v1.Pod) map[string][]topologymana
 				continue
 			}
 			klog.InfoS("Regenerating TopologyHints for resource already allocated to pod", "resource", resource, "pod", klog.KObj(pod))
-			deviceHints[resource] = m.generateDeviceTopologyHints(resource, allocated, sets.String{}, requested)
+			deviceHints[resource] = m.generateDeviceTopologyHints(resource, allocated, sets.Set[string]{}, requested)
 			continue
 		}
 
@@ -132,7 +132,7 @@ func (m *ManagerImpl) GetPodTopologyHints(pod *v1.Pod) map[string][]topologymana
 
 		// Generate TopologyHints for this resource given the current
 		// request size and the list of available devices.
-		deviceHints[resource] = m.generateDeviceTopologyHints(resource, available, sets.String{}, requested)
+		deviceHints[resource] = m.generateDeviceTopologyHints(resource, available, sets.Set[string]{}, requested)
 	}
 
 	return deviceHints
@@ -148,12 +148,12 @@ func (m *ManagerImpl) deviceHasTopologyAlignment(resource string) bool {
 	return false
 }
 
-func (m *ManagerImpl) getAvailableDevices(resource string) sets.String {
+func (m *ManagerImpl) getAvailableDevices(resource string) sets.Set[string] {
 	// Strip all devices in use from the list of healthy ones.
 	return m.healthyDevices[resource].Difference(m.allocatedDevices[resource])
 }
 
-func (m *ManagerImpl) generateDeviceTopologyHints(resource string, available sets.String, reusable sets.String, request int) []topologymanager.TopologyHint {
+func (m *ManagerImpl) generateDeviceTopologyHints(resource string, available sets.Set[string], reusable sets.Set[string], request int) []topologymanager.TopologyHint {
 	// Initialize minAffinitySize to include all NUMA Nodes
 	minAffinitySize := len(m.numaNodes)
 
diff --git a/pkg/kubelet/cm/devicemanager/topology_hints_test.go b/pkg/kubelet/cm/devicemanager/topology_hints_test.go
index a7bc5157366..43bfafe373c 100644
--- a/pkg/kubelet/cm/devicemanager/topology_hints_test.go
+++ b/pkg/kubelet/cm/devicemanager/topology_hints_test.go
@@ -61,8 +61,8 @@ func TestGetTopologyHints(t *testing.T) {
 	for _, tc := range tcases {
 		m := ManagerImpl{
 			allDevices:       NewResourceDeviceInstances(),
-			healthyDevices:   make(map[string]sets.String),
-			allocatedDevices: make(map[string]sets.String),
+			healthyDevices:   make(map[string]sets.Set[string]),
+			allocatedDevices: make(map[string]sets.Set[string]),
 			podDevices:       newPodDevices(),
 			sourcesReady:     &sourcesReadyStub{},
 			activePods:       func() []*v1.Pod { return []*v1.Pod{tc.pod} },
@@ -71,7 +71,7 @@ func TestGetTopologyHints(t *testing.T) {
 
 		for r := range tc.devices {
 			m.allDevices[r] = make(DeviceInstances)
-			m.healthyDevices[r] = sets.NewString()
+			m.healthyDevices[r] = sets.New[string]()
 
 			for _, d := range tc.devices[r] {
 				m.allDevices[r][d.ID] = d
@@ -84,7 +84,7 @@ func TestGetTopologyHints(t *testing.T) {
 				for r, devices := range tc.allocatedDevices[p][c] {
 					m.podDevices.insert(p, c, r, constructDevices(devices), nil)
 
-					m.allocatedDevices[r] = sets.NewString()
+					m.allocatedDevices[r] = sets.New[string]()
 					for _, d := range devices {
 						m.allocatedDevices[r].Insert(d)
 					}
@@ -414,8 +414,8 @@ func TestTopologyAlignedAllocation(t *testing.T) {
 	for _, tc := range tcases {
 		m := ManagerImpl{
 			allDevices:            NewResourceDeviceInstances(),
-			healthyDevices:        make(map[string]sets.String),
-			allocatedDevices:      make(map[string]sets.String),
+			healthyDevices:        make(map[string]sets.Set[string]),
+			allocatedDevices:      make(map[string]sets.Set[string]),
 			endpoints:             make(map[string]endpointInfo),
 			podDevices:            newPodDevices(),
 			sourcesReady:          &sourcesReadyStub{},
@@ -424,7 +424,7 @@ func TestTopologyAlignedAllocation(t *testing.T) {
 		}
 
 		m.allDevices[tc.resource] = make(DeviceInstances)
-		m.healthyDevices[tc.resource] = sets.NewString()
+		m.healthyDevices[tc.resource] = sets.New[string]()
 		m.endpoints[tc.resource] = endpointInfo{}
 
 		for _, d := range tc.devices {
@@ -441,7 +441,7 @@ func TestTopologyAlignedAllocation(t *testing.T) {
 			}
 		}
 
-		allocated, err := m.devicesToAllocate("podUID", "containerName", tc.resource, tc.request, sets.NewString())
+		allocated, err := m.devicesToAllocate("podUID", "containerName", tc.resource, tc.request, sets.New[string]())
 		if err != nil {
 			t.Errorf("Unexpected error: %v", err)
 			continue
@@ -603,8 +603,8 @@ func TestGetPreferredAllocationParameters(t *testing.T) {
 	for _, tc := range tcases {
 		m := ManagerImpl{
 			allDevices:            NewResourceDeviceInstances(),
-			healthyDevices:        make(map[string]sets.String),
-			allocatedDevices:      make(map[string]sets.String),
+			healthyDevices:        make(map[string]sets.Set[string]),
+			allocatedDevices:      make(map[string]sets.Set[string]),
 			endpoints:             make(map[string]endpointInfo),
 			podDevices:            newPodDevices(),
 			sourcesReady:          &sourcesReadyStub{},
@@ -613,13 +613,13 @@ func TestGetPreferredAllocationParameters(t *testing.T) {
 		}
 
 		m.allDevices[tc.resource] = make(DeviceInstances)
-		m.healthyDevices[tc.resource] = sets.NewString()
+		m.healthyDevices[tc.resource] = sets.New[string]()
 		for _, d := range tc.allDevices {
 			m.allDevices[tc.resource][d.ID] = d
 			m.healthyDevices[tc.resource].Insert(d.ID)
 		}
 
-		m.allocatedDevices[tc.resource] = sets.NewString()
+		m.allocatedDevices[tc.resource] = sets.New[string]()
 		for _, d := range tc.allocatedDevices {
 			m.allocatedDevices[tc.resource].Insert(d)
 		}
@@ -639,17 +639,17 @@ func TestGetPreferredAllocationParameters(t *testing.T) {
 			opts: &pluginapi.DevicePluginOptions{GetPreferredAllocationAvailable: true},
 		}
 
-		_, err := m.devicesToAllocate("podUID", "containerName", tc.resource, tc.request, sets.NewString(tc.reusableDevices...))
+		_, err := m.devicesToAllocate("podUID", "containerName", tc.resource, tc.request, sets.New[string](tc.reusableDevices...))
 		if err != nil {
 			t.Errorf("Unexpected error: %v", err)
 			continue
 		}
 
-		if !sets.NewString(actualAvailable...).Equal(sets.NewString(tc.expectedAvailable...)) {
+		if !sets.New[string](actualAvailable...).Equal(sets.New[string](tc.expectedAvailable...)) {
 			t.Errorf("%v. expected available: %v but got: %v", tc.description, tc.expectedAvailable, actualAvailable)
 		}
 
-		if !sets.NewString(actualAvailable...).Equal(sets.NewString(tc.expectedAvailable...)) {
+		if !sets.New[string](actualAvailable...).Equal(sets.New[string](tc.expectedAvailable...)) {
 			t.Errorf("%v. expected mustInclude: %v but got: %v", tc.description, tc.expectedMustInclude, actualMustInclude)
 		}
 
@@ -903,11 +903,11 @@ func TestGetPodDeviceRequest(t *testing.T) {
 
 	for _, tc := range tcases {
 		m := ManagerImpl{
-			healthyDevices: make(map[string]sets.String),
+			healthyDevices: make(map[string]sets.Set[string]),
 		}
 
 		for _, res := range tc.registeredDevices {
-			m.healthyDevices[res] = sets.NewString()
+			m.healthyDevices[res] = sets.New[string]()
 		}
 
 		accumulatedResourceRequests := m.getPodDeviceRequest(tc.pod)
@@ -925,8 +925,8 @@ func TestGetPodTopologyHints(t *testing.T) {
 	for _, tc := range tcases {
 		m := ManagerImpl{
 			allDevices:       NewResourceDeviceInstances(),
-			healthyDevices:   make(map[string]sets.String),
-			allocatedDevices: make(map[string]sets.String),
+			healthyDevices:   make(map[string]sets.Set[string]),
+			allocatedDevices: make(map[string]sets.Set[string]),
 			podDevices:       newPodDevices(),
 			sourcesReady:     &sourcesReadyStub{},
 			activePods:       func() []*v1.Pod { return []*v1.Pod{tc.pod, {ObjectMeta: metav1.ObjectMeta{UID: "fakeOtherPod"}}} },
@@ -935,7 +935,7 @@ func TestGetPodTopologyHints(t *testing.T) {
 
 		for r := range tc.devices {
 			m.allDevices[r] = make(DeviceInstances)
-			m.healthyDevices[r] = sets.NewString()
+			m.healthyDevices[r] = sets.New[string]()
 
 			for _, d := range tc.devices[r] {
 				//add `pluginapi.Device` with Topology
@@ -949,7 +949,7 @@ func TestGetPodTopologyHints(t *testing.T) {
 				for r, devices := range tc.allocatedDevices[p][c] {
 					m.podDevices.insert(p, c, r, constructDevices(devices), nil)
 
-					m.allocatedDevices[r] = sets.NewString()
+					m.allocatedDevices[r] = sets.New[string]()
 					for _, d := range devices {
 						m.allocatedDevices[r].Insert(d)
 					}
diff --git a/pkg/kubelet/cm/devicemanager/types.go b/pkg/kubelet/cm/devicemanager/types.go
index fb330568adc..7e3261c667d 100644
--- a/pkg/kubelet/cm/devicemanager/types.go
+++ b/pkg/kubelet/cm/devicemanager/types.go
@@ -33,7 +33,7 @@ import (
 // Manager manages all the Device Plugins running on a node.
 type Manager interface {
 	// Start starts device plugin registration service.
-	Start(activePods ActivePodsFunc, sourcesReady config.SourcesReady, initialContainers containermap.ContainerMap, initialContainerRunningSet sets.String) error
+	Start(activePods ActivePodsFunc, sourcesReady config.SourcesReady, initialContainers containermap.ContainerMap, initialContainerRunningSet sets.Set[string]) error
 
 	// Allocate configures and assigns devices to a container in a pod. From
 	// the requested device resources, Allocate will communicate with the
diff --git a/pkg/kubelet/cm/helpers.go b/pkg/kubelet/cm/helpers.go
index 6be3e272307..d1c0beabf74 100644
--- a/pkg/kubelet/cm/helpers.go
+++ b/pkg/kubelet/cm/helpers.go
@@ -52,14 +52,14 @@ func hardEvictionReservation(thresholds []evictionapi.Threshold, capacity v1.Res
 	return ret
 }
 
-func buildContainerMapAndRunningSetFromRuntime(ctx context.Context, runtimeService internalapi.RuntimeService) (containermap.ContainerMap, sets.String) {
+func buildContainerMapAndRunningSetFromRuntime(ctx context.Context, runtimeService internalapi.RuntimeService) (containermap.ContainerMap, sets.Set[string]) {
 	podSandboxMap := make(map[string]string)
 	podSandboxList, _ := runtimeService.ListPodSandbox(ctx, nil)
 	for _, p := range podSandboxList {
 		podSandboxMap[p.Id] = p.Metadata.Uid
 	}
 
-	runningSet := sets.NewString()
+	runningSet := sets.New[string]()
 	containerMap := containermap.NewContainerMap()
 	containerList, _ := runtimeService.ListContainers(ctx, nil)
 	for _, c := range containerList {
diff --git a/pkg/kubelet/cm/internal_container_lifecycle_linux.go b/pkg/kubelet/cm/internal_container_lifecycle_linux.go
index cb7c0cfa543..0c3bb2e4999 100644
--- a/pkg/kubelet/cm/internal_container_lifecycle_linux.go
+++ b/pkg/kubelet/cm/internal_container_lifecycle_linux.go
@@ -23,7 +23,8 @@ import (
 	"strconv"
 	"strings"
 
-	"k8s.io/api/core/v1"
+	v1 "k8s.io/api/core/v1"
+	"k8s.io/apimachinery/pkg/util/sets"
 	runtimeapi "k8s.io/cri-api/pkg/apis/runtime/v1"
 )
 
@@ -39,7 +40,7 @@ func (i *internalContainerLifecycleImpl) PreCreateContainer(pod *v1.Pod, contain
 		numaNodes := i.memoryManager.GetMemoryNUMANodes(pod, container)
 		if numaNodes.Len() > 0 {
 			var affinity []string
-			for _, numaNode := range numaNodes.List() {
+			for _, numaNode := range sets.List(numaNodes) {
 				affinity = append(affinity, strconv.Itoa(numaNode))
 			}
 			containerConfig.Linux.Resources.CpusetMems = strings.Join(affinity, ",")
diff --git a/pkg/kubelet/cm/memorymanager/fake_memory_manager.go b/pkg/kubelet/cm/memorymanager/fake_memory_manager.go
index 364614502d4..46874e50050 100644
--- a/pkg/kubelet/cm/memorymanager/fake_memory_manager.go
+++ b/pkg/kubelet/cm/memorymanager/fake_memory_manager.go
@@ -50,7 +50,7 @@ func (m *fakeManager) AddContainer(pod *v1.Pod, container *v1.Container, contain
 	klog.InfoS("Add container", "pod", klog.KObj(pod), "containerName", container.Name, "containerID", containerID)
 }
 
-func (m *fakeManager) GetMemoryNUMANodes(pod *v1.Pod, container *v1.Container) sets.Int {
+func (m *fakeManager) GetMemoryNUMANodes(pod *v1.Pod, container *v1.Container) sets.Set[int] {
 	klog.InfoS("Get MemoryNUMANodes", "pod", klog.KObj(pod), "containerName", container.Name)
 	return nil
 }
diff --git a/pkg/kubelet/cm/memorymanager/memory_manager.go b/pkg/kubelet/cm/memorymanager/memory_manager.go
index b8c55b74597..04cf2d6f533 100644
--- a/pkg/kubelet/cm/memorymanager/memory_manager.go
+++ b/pkg/kubelet/cm/memorymanager/memory_manager.go
@@ -83,7 +83,7 @@ type Manager interface {
 	GetPodTopologyHints(*v1.Pod) map[string][]topologymanager.TopologyHint
 
 	// GetMemoryNUMANodes provides NUMA nodes that are used to allocate the container memory
-	GetMemoryNUMANodes(pod *v1.Pod, container *v1.Container) sets.Int
+	GetMemoryNUMANodes(pod *v1.Pod, container *v1.Container) sets.Set[int]
 
 	// GetAllocatableMemory returns the amount of allocatable memory for each NUMA node
 	GetAllocatableMemory() []state.Block
@@ -213,9 +213,9 @@ func (m *manager) AddContainer(pod *v1.Pod, container *v1.Container, containerID
 }
 
 // GetMemoryNUMANodes provides NUMA nodes that used to allocate the container memory
-func (m *manager) GetMemoryNUMANodes(pod *v1.Pod, container *v1.Container) sets.Int {
+func (m *manager) GetMemoryNUMANodes(pod *v1.Pod, container *v1.Container) sets.Set[int] {
 	// Get NUMA node affinity of blocks assigned to the container during Allocate()
-	numaNodes := sets.NewInt()
+	numaNodes := sets.New[int]()
 	for _, block := range m.state.GetMemoryBlocks(string(pod.UID), container.Name) {
 		for _, nodeID := range block.NUMAAffinity {
 			// avoid nodes duplication when hugepages and memory blocks pinned to the same NUMA node
diff --git a/pkg/kubelet/cm/topologymanager/policy_options.go b/pkg/kubelet/cm/topologymanager/policy_options.go
index 15f94c696d2..1ef211cd0bd 100644
--- a/pkg/kubelet/cm/topologymanager/policy_options.go
+++ b/pkg/kubelet/cm/topologymanager/policy_options.go
@@ -30,11 +30,11 @@ const (
 )
 
 var (
-	alphaOptions = sets.NewString()
-	betaOptions  = sets.NewString(
+	alphaOptions = sets.New[string]()
+	betaOptions  = sets.New[string](
 		PreferClosestNUMANodes,
 	)
-	stableOptions = sets.NewString()
+	stableOptions = sets.New[string]()
 )
 
 func CheckPolicyOptionAvailable(option string) error {
diff --git a/pkg/kubelet/cm/topologymanager/policy_options_test.go b/pkg/kubelet/cm/topologymanager/policy_options_test.go
index d01b63db799..c9872ec76b9 100644
--- a/pkg/kubelet/cm/topologymanager/policy_options_test.go
+++ b/pkg/kubelet/cm/topologymanager/policy_options_test.go
@@ -112,7 +112,7 @@ func TestNewTopologyManagerOptions(t *testing.T) {
 	}
 
 	betaOptions.Insert(fancyBetaOption)
-	alphaOptions = sets.NewString(fancyAlphaOption)
+	alphaOptions = sets.New[string](fancyAlphaOption)
 
 	for _, tcase := range testCases {
 		t.Run(tcase.description, func(t *testing.T) {