The Availability of a Cluster with the Protection of Real-Time Data Replication
No Shared Disk Storage
Today, companies are challenged to answer customer demands quickly, accurately and also to provide continuous information. Data availability and protection are crucial to business success. GeoCluster gives MSCS clusters increased infrastructure protection and higher levels of availability than the standard shared storage configuration.
Servers in an MSCS configuration share one storage sub system and must be inclose proximity to each other, anywhere from a few feet if connected via SCSI, to no more than several kilometers if connected by Fibre Channel. The computers, or nodes, take turns accessing the shared storage, which is a single point of failure that leaves the entire cluster vulnerable to hardware related or geographic disasters. With GeoCluster, each server has its own data copy, and geographic limitations are eliminated. It is a practical and cost-effective disaster-protection solution for network administrators.
By integrating its award-winning replication technology with MSCS capabilities, NSI Software provides seamless configuration, control and administration of data replication within a cluster. Replication is real time; it occurs continuously over any shared or private IP-based LAN, VLAN or SAN connection. Data is captured as changes are made, and as a result, data integrity is always preserved. In fact, replicated data is protected from accidental modification, with inactive nodes kept in read-only status.
How does it work?
GeoCluster uses local copies of clustered data stored on each node, eliminating a single point of failure. Data is synchronized using standard TCP/I P connections. That means servers can be located almost anywhere. For example, what if your primary node is headquarterd in the city and your back up is located in a remote suburban office? No need to worry. GeoCluster keeps your business safe from virtually any distance.
How GeoCluster Works
- Network users access cluster resources using location independent virtual server identities.
- Cluster nodes maintain local copies of clustered data that is replicated over IP connections.
- The GeoCluster Replicated Disk resource controls data replication and flow, as well as access to disk contents.
- Read-only access to replicated disks on the inactive node ensures data integrity.
- A node failure or manual intervention moves a virtual server and all of its resources to a remaining cluster node. Users continue to access the virtual server identity and replicated data.
- The cluster quorum log is also replicated to all nodes in the cluster simply by designating a replicated disk as the cluster quorum resource.
- Quorum ownership is determined by a node’s ability to access arbitration shares on other machines over the network.
To achieve the highest possible availability, NSI Software knows that business-critical applications must be cluster aware. High-level applications, such as Microsoft Exchange and SQL Server, run on Geo Cluster in the same manner they run on MSCS. Even database or web site-level functions are supported by GeoCluster, just as they are on a shared disk cluster.
Data is moved safely and simply from a failed server to an alternate location, by design. With GeoCluster, there is no need for risky, unapproved techniques like registry key and DLL copying. Individual databases can be moved independently between nodes as needed. In short, GeoCluster’s data replication is transparent to applications running on an MSCS cluster.
Does your company need amultitude of options and flexibility when it comes to data availability? Most do. And that’s why GeoCluster is the right solution for so many companies.
By using Microft’s MSCS API to provide application fail over, each individual application can be monitored and automatically restarted if a failure occurs. Group failover allows an entire application and all of its supporting functions to be moved to another node or to to be manually relocated to balance application loads . Of course, GeoCluster determines whether a node is operational, but it operates under a much larger purposeto ensure a higher level of data availability when compared with other cluster technologies.
A disaster occurs and your system fails over to a remote location. When your primary node is up and running again, how do you reverse the process and resume normal operations? Multiple servers can protect each other in an active/active configuration, and MSCS virtual serves make network identities and services independent of location. With GeoCluster, failback to the original node is as seamless as failover.
Features and Benefits
Improved Data Availability
GeoCluster uses byte-level replication, which ensures maximum application performance and minimum bandwidth requirements. It reduces exposure to data loss and eliminates the single point of failure of shared disk storage through data redundancy.
Application failover is fast and automatic, eliminating or greatly reducing any downtime. Even in a disastrous situation, your business can operate continuously. GeoCluster has an intricate checks-and-balances allowing MSCS cluster-aware applications to be monitored by and interact directly with the cluster service.
Disaster Recovery Protection
Replicated disks are synchronized over any IP-based LAN, VLAN or SAN connections, allowing up-to-the-minute offsite date protection and failover. GeoCluster even works efficiently over slower long distance links, so business-critical information is always available. It’s high-level protection without geographic limitations.
Double-Take Management Framework
GeoCluster is part of the Double-Take Management Framework, which provides consistent data recovery and high availability solutions for Windows 2000/NT, Solaris, and NetWare. Robust event and statistic monitoring is available through Double-Take utilities and the Windows Performance Monitor and Event Viewer. In addition, GeoCluster uses the Double-Take SNMP MIB to provide extensive statistics monitoring and alerting to any standard SNMP-based management system. All Double-Take and GeoCluster products can be managed by the same GUI, text-based or command line clients running on supported platforms locally or remotely attached. GeoCluster can be used with or without the standard Double-Take product. When used in conjunction with Double-Take, GeoCluster provides additional data protection options, replicates data outside a single cluster, and takes advantage of many other capabilities that are not provided when used alone, such as cross-platform replication, data migration, and bandwidth limiting.
GeoCluster users guide
- GeoCluster is no longer available from AMT Software.
GeoCluster requires Microsoft Cluster Service 1.0 or greater (included with Microsoft Windows NT Server 4.0 Enterprise Edition, Windows 2000 Advanced Server and Windows 2000 DataCenter Server) and a TCP/IP connection between nodes on the same logical IP subnet (using direct LAN connection or VLAN technology), with maximum round trip latency between nodes of no more than a half a second. Multiple network links are recommended to isolate public and private traffic. Cluster arbitration requires each node to have full access to at least one independent machine on the network.
Intel-based hardware is required for GeoCluster. Each node should have matching drive letters and the same amount of disk space. Maintaining the quorum database log on a separate drive letter from application data volumes is also recommended. The Windows 2000/NT boot volume cannot be replicated. Microsoft support for MSCS and MSCS-based Microsoft applications requires that the cluster configuration appear on the Microsoft Hardware Compatibility List under category “Cluster.”
Nearly any MSCS-supported application can be configured using GeoCluster. MSCS also supports generic applications and services.