Scale Computing for Virtualization
Virtualization is a fast-growing technology, and with reason. Virtualizers like VMware enable IT to leverage servers, storage, and networking for a much-improved computing infrastructure. But virtualization comes with its own set of challenges, especially around storage systems. These systems house data from dozens (or more) of virtual servers and must manage I/O from each server’s network interface card (NIC). This demanding environment causes spikes in workload, performance, and capacity demands that challenge IT and the storage infrastructure.
Scale responds with its ICS cluster architecture, which provides grid-based performance for large and unpredictable virtualized workloads. These advanced but highly cost-effective storage systems are extremely scalable, growing node by node with no disruption to data or services. ICS is high performance and highly available, with full redundancy and data replication built-in.
ICS clusters are economical and powerful from the start. The entry-level 3TB Starter Cluster starts with three S1’s with 1TB of useable capacity. The S2 comes in at 2 useable TB and the S4 at 4 useable TB. To observe VMware’s data protection best practices, the nodes are highly redundant and automatically replicated. If a single node fails, data is immediately re-replicated and another node seamlessly takes the load. ICS also preserves fast data access even during spiking workloads, and maintains the highest levels of system uptime.
Benefits
- IT can efficiently access system-wide management tools from each node
- Observes VMware best practices for storage
- Extremely high system uptime preserves storage availability to the virtualized network
- Data and service protection operate even during drive failure
- Scaling is a simple matter of adding additional storage nodes; no downtime or service disruption
- Grid architecture preserves data access speeds and easily manages spiking workloads
Scale Computing Tames the Virtualized Environment
Server virtualization allows users to run multiple virtual machines (VMs) on a single physical server. Virtualization works by inserting a software layer called a hypervisor that administers the guest machines and allocates the physical server’s hard resources to its guests. Each VM is capable of running its own operating system and applications, which makes virtualization an extremely efficient computing environment. Advantages include:
- Recover fast from data disasters. The efficient virtualization architecture simplifies data protection like snapshots. Recovery is fast in the event of virus attacks or network outages.
- Reduce management costs by up to 300%. Virtualization vastly reduces the number of servers that IT must manage.
- Significant energy savings. By consolidating and retiring servers, IT can save dramatically on power and cooling costs.
- Avoid data center build out. Virtual servers take up far less real estate. This drastically reduces equipment footprints and extends the life of the data center.
- Leverage legacy applications. Lengthen the life of legacy applications by operating them on an efficient VM.
If Virtualization Is so Great, What’s the Problem?
The problem is that virtualization makes it easy to run more and more applications and to generate more and more snapshots. Where the older data center might have run 5 application servers, now they’re using the same 5 physical servers to host dozens of applications. The result of so much data is a severe strain on storage performance and capacity. Storage performance suffers when application data and virtual snapshots bottleneck at the storage controllers, and the sheer volume of data threatens to overwhelm storage capacity.
If IT uses direct-attached storage (DAS) for each physical server, then they must over-provision each storage device to keep it from crossing capacity thresholds. And there was no way to load balance across multiple DAS, meaning that one DAS might be close to capacity while another DAS is consistently under-used. Even if IT is using a SAN or NAS, the same problems apply. There would be more centralized capacity but the same performance issues. And scalability would still be a serious problem: in a large virtualized environment, downing networked storage to add capacity and upgrades is a downtime nightmare.
Enter Scale ICS
Scale Computing solves these problems and more. Older storage architectures have serious provisioning, performance and capacity issues in the virtualized environment. In contrast, Scale’s advanced but highly cost-effective storage systems are extremely scalable, growing node by node up to petabyte levels — with no disruption to data or services. Scale ICS streamlines large data movement from virtual servers and seamlessly stores them across multiple nodes. ICS is high performance and highly available, with built-in hardware redundancy. It generates its own data replication and efficient snapshots for robust and efficient data protection.
You might expect this enterprise-level of power and performance to come with a matching enterprise-level price tag. But ICS clusters are economical and powerful from the start. The entry-level 3TB Starter Cluster starts with three S1’s with 1TB of useable capacity. No single point of failure, no bottlenecking storage controller. Once customers have the 3-node cluster in place, they may grow their Scale clusters at will by single nodes or more. Customers can also start off with a bigger bang: the S2 comes in at 2 useable TB and the S4 at 4 useable TB.
If a single node fails, data is immediately re-replicated and another node seamlessly takes the load. ICS preserves fast data access even during spiking workloads, and maintains the highest levels of system uptime in highly demanding virtualized environments.
Use Case
Top virtualization maker VMware provides virtualization as a service with its vCloud product. vCloud allows remote customers to create virtual machines as needed and add compute capacity using an online interface. Service provider BlueLock uses vCloud to remotely host virtualization services for its customers. This multi-customer environment generates a lot of data and generates it fast, and must have iron-clad performance, capacity and protection – all at an affordable price.
That’s why BlueLock chose Scale Computing. Scale’s clustered and parallel architecture seamlessly scales to multi-petabytes of high performance and highly available storage. And since it runs on commodity hardware, it can offer an extensive set of enterprise-class features at an extremely reasonable price.
Scale Features and Benefits
| Feature | Benefit |
|---|---|
| System-wide management tools are available from each node | Storage system management is flexible and efficient |
| Observes VMware best practices for storage | No more storage bottlenecks in virtualized environments |
| Extremely high system uptime | Enables high availability in the most demanding virtualized environment |
| Scaling is a simple matter of adding additional storage nodes | Scales from 3TB up to 2PB with no downtime or service disruption |
| Clustered and parallel grid architecture | Preserves data access speeds and high performance levels |
| Copyright © 2010 Scale Computing