LISP Stretched Subnet Mode for Data Center Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data center migrations often require changing IP addresses, subnet configurations, and default gateways, which is cumbersome and costly, and existing migration solutions can be unreliable and complex, especially when trying to maintain IP address consistency across locations.
Innovation Solution
The use of Locator/Identifier Separation Protocol (LISP) in stretched subnet mode allows for migration of network elements without changing their IP addresses, subnet masks, or default gateways by separating identity and location through Routing Locators and Endpoint Identifiers, enabling Layer 3 extensions across data centers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional IP address reconfiguration is used during data center migration, then network connectivity can be established at the new location, but the migration process becomes cumbersome and costly due to changing IP addresses, subnet configurations, and default gateways
Solution Approach 1:
The patent segments the IP address into two separate components: Endpoint Identifier (EID) which remains constant and Locator Identifier (LID) which changes with location. This segmentation allows the server to maintain its identity (EID) while migrating to new locations with different LIDs, eliminating the need to reconfigure the entire IP address and associated network settings during migration.
Solution Approach 2:
The patent introduces LISP routers and a mapping database as intermediaries between the migrating server and the network. The mapping database stores the mapping between EID and LID, allowing seamless translation of network addresses during migration without requiring changes to the server's configuration, thus simplifying the migration process while maintaining connectivity.
2Adaptability or versatility
If IP addresses are changed during data center migration, then servers can be deployed at new locations, but the complexity and cost of reconfiguring network settings increases
Solution Approach 1:
By separating the IP address into EID (identity) and LID (location) components, the patent enables servers to be deployed at multiple locations without increasing configuration complexity. The EID remains constant while LIDs are automatically updated in the mapping database, allowing versatile deployment without additional reconfiguration effort.
Solution Approach 2:
The LISP routing system automatically updates the mapping database with new LID-EID mappings when servers migrate to new locations. This self-service mechanism eliminates manual reconfiguration of network settings, reducing both the complexity and time required for server deployment at new locations.
3Stability of the object's composition
If Layer 2 stretching is used to maintain IP addresses during migration, then IP address consistency is preserved, but the solution becomes unreliable due to broadcast domain interconnection issues
Solution Approach 1:
Instead of directly stretching Layer 2 broadcast domains, the patent introduces LISP routers and mapping database as intermediaries at Layer 3. These intermediaries maintain IP address consistency through automated EID-LID mapping while avoiding the reliability issues of Layer 2 stretching by using encapsulated routing at Layer 3.
Solution Approach 2:
The patent transitions from Layer 2 broadcast domain stretching to Layer 3 encapsulated routing with mapping database. This dimensional shift from data link layer to network layer provides the same IP address consistency benefit while avoiding the reliability problems inherent in Layer 2 approaches.
4Productivity
If full subnet migration is performed, then all servers can be moved to the new data center, but the data transfer requirements and migration time increase significantly
Solution Approach 1:
The patent enables incremental migration by allowing individual servers or small groups of servers to be migrated independently. Each server maintains its EID while only its LID needs to be updated in the mapping database, enabling phased migration without requiring migration of entire subnets simultaneously, thus reducing data transfer requirements and migration time.
Solution Approach 2:
Instead of migrating complete subnets, the patent allows migration of individual servers or partial groups. The LISP routing system handles partial migrations efficiently by updating only the necessary EID-LID mappings in the database, reducing the scope of data transfer and migration time while maintaining full network functionality.
Data Source
AI summary
The present disclosure describes methods and systems for enabling a migration of network elements from a first location to a second location remote from the first location without changing the Internet Protocol (IP) addresses, subnet mask, and/or default gateway of the network elements. The first location has a first Locator/Identifier Separation Protocol (LISP) router configured on a stick and the second location having a second LISP router configured on a stick. Both the first LISP router and the second LISP router are on the same subnet. Effectively, LISP provides a Layer 3 extension stretching a subnet across the first location and the second location (Stretched Subnet Mode (SSM)). By implementing LISP routers in this manner, system engineers can migrate network elements easily between two locations.