In the first part of my blog series, I shed light on the key trends that impact the next-generation data center including consolidation, virtualization, connectivity, convergence and green IT. In this part, it’s time to discuss what are the best-of-breed solutions to address these trends in the most cost-effective way. I chose to start with data center consolidation and virtualization not only because I believe they’re the prominent ones – but also since they go hand in hand together. So here we go:
Data center consolidation – There’s a clear economic incentive behind data center consolidation: reducing the number of data centers, the number of servers and network infrastructure means less CAPEX (product cost, cabling, closets, etc.) and OPEX (data center space, electricity, cooling, labor and more). While organizations historically implemented consolidation projects by simply using stronger servers located in less geographical locations, today’s common practice is to use virtualization technologies (which we’ll discuss in the next section). But fewer data centers, fewer servers and more “flat” network means putting your eggs in fewer baskets, meaning that every layer in the data center is becoming much more critical – and therefore needs to be more resilient. Therefore, to ensure continuous business continuity, it is essential to use solutions that ensure high-availability (HA) and that are fault-tolerant (in fact, today IT departments are also looking for “self-healing” grade solutions). Consolidation is not limited to the server domain; it happens also in switches, for example, today’s next-generation switches connect both application and storage by leveraging technologies such as Fibre Channel over Ethernet (FCoE) and unified ports (these are relevant also for convergence which we’ll discuss later on). So what about the ADC?
From an application delivery controller (ADC) viewpoint, consolidation is more relevant than ever: it means taking many legacy ADC units and consolidating them into a stronger, resilient, highly-performing and consolidated platform – such as Radware Alteon VX. To ensure end-to-end application SLA, it is vital to ensure “cross-site high-availability,” i.e. to ensure undisrupted service even in case of a complete data center failure. Using Global Server Load Balancing (GSLB) solution delivers fastest failover (from the active site to the backup site) while ensuring transaction completion, at all times. Finally, data center consolation also implies using enterprise-grade solution to continually backup and replicate business-critical data, as well as to address more complex regulatory compliance via audit logs, reports and analytics.
Virtualization – Virtualization is a common practice to consolidate hardware by running multiple, logical entities on fewer physical hardware – therefore it is considered the “beating heart” of consolidation. But it is capable of delivering much more. By using management orchestration systems (such as VMware vCenter Orchestrator, IBM Tivoli, etc.), the IT manager can easily support operations that were considered to be very complex, costly and risky. Data center dynamic changes including capacity growth, new services, topology changes or configuration refinements can be addressed in minutes and without any IT equipment changes/replacements – which increases the business agility.
In addition, virtualization allows a more efficient data center. For starters, complex tasks can be fully automated to reduce human-intervention, error and delay. Moreover, in many cases servers are underutilized during non-peak times, and using advanced capacity measurement and dynamic allocation introduced by virtualization management systems can increase data center/network elasticity. For example, unused capacity can be allocated to another service that requires them at a given time, and re-allocated ad-hoc when needed.
From an ADC viewpoint, today’s advanced ADC solutions feature built-in hypervisors (such as Radware’s ADC-VX™ employed in the Alteon platforms) that enable to virtualize ADC services into a virtual ADC (vADC) just like servers can be virtualized using server virtualization hypervisors. Here, scalability and density is what matters: the higher your ADC solution can scale (in terms of throughput, services and # of vADC instances) – the better the ADC can address data center dynamic changes. For instance, Radware Alteon 10000 can scale up to 80Gbps and 256 vADCs – which means that once you deploy the solution, you can literally forget about changing the device! (You can scale it using software licenses only.)
Using a solution that lets you deploy ADC in different form factors (physical, virtualized/multi-tenant and software) gives you the freedom to scale and automate your IT operations. Here, the sky is the limit, and you can consider the ADC not as a bunch of hardware unit but rather an “ADC Fabric” (connecting the network fabric and the compute fabric) – which delivers a set of scalable, resilient application delivery services. And the best part: using virtualization management plug-in and SDK (such as Radware’s vDirect), the ADC can play a role part of the data center automation so that automatic process workflow include layer 4-7 policies and configuration changes.
In part 3 of this blog, we’ll discuss the rest of the trends – connectivity, convergence and green IT – and see what they mean from a specific application delivery viewpoint.
Until next time,