Cisco’s new data-center architecture comprises compute, network, storage access and virtualization resources in a single rackable system designed to cut IT infrastructure costs and complexity, stretch existing IT investments and allow enterprise customers to build an agile data center that they can easily extend for future growth, according to the company.
Taking a step into the server market, where it will compete with long-standing partners like Hewlett-Packard and IBM, Cisco announced the UCS B-Series blades, based on upcoming Intel Nehalem processors. Cisco said the blades incorporate extended memory technology for applications with large data sets.
However, Cisco also is teaming up with software partners such as Microsoft, VMware and BMC to provide technology for its new system. Cisco will prepackage, resell and support Windows Server 2003, Windows Server 2008 with Hyper-V technology and Microsoft SQL Server 2008 as part of the Unified Computing System. For its part, VMware is providing virtualization technology to the new system and BMC is contributing resource-management software.
For the Linux crowd, Cisco also will sell and support Red Hat Enterprise Linux as of the new system and will support Red Hat’s forthcoming virtualization portfolio as well. Like its new rivals HP and IBM, Cisco is becoming an OEM (original equipment manufacturer) partner to server OS vendors, selling their software on its new blades.
Systems integrator Accenture has already signed on to be a services partner, introducing on Monday four services options for its customers to deploy the Unified Computing System. Cisco also is inviting its broad network of channel partners to work with the company to provide the new infrastructure to enterprise customers.
Prior to Monday’s announcement, Cisco had laid out its intention to eliminate the manual integration of computing and storage platforms with networks and virtualization systems. In a recent blog posting, CTO Padmasree Warrior acknowledged this will lead Cisco to compete with some of its partners.
Most enterprises have been able to realize the benefits of the first phase of virtualization, consolidating their data centers for economies of scale and simpler management. But it’s been hard to achieve full virtualization, in which virtual machines can continuously move among servers, analysts say. This would allow for adding processing power as demand for an application grows, or for moving tasks off a physical server at night for hardware maintenance.
Vendors of servers, storage and software all can play roles in managing resources in virtualized data centers. Cisco executives have said the network is the best place to tackle many of these tasks because it is the only element of the infrastructure that touches everything.
IBM and HP have not overlooked the importance of networks in controlling data centers. IBM has aligned with Cisco rival Juniper Networks in a broad strategy called Stratus, and HP is expected to increasingly tie its growing ProCurve networking business in with its computing offerings.
Cisco has pushed to bring more functions into the network infrastructure for several years. These have included security, application optimization, and adaptation of multimedia to fit different clients and purposes. As revenue growth from its core routing and switching businesses slows, the company is aggressively branching out into new areas, including consumer electronics and the Telepresence high-definition videoconferencing line.
Note: When you purchase something after clicking links in our articles, we may earn a small commission. Read ouraffiliate link policyfor more details.