Riverbed Technology recently launched a new version of its product – Riverbed RiOS 5.0. According to Apurva Dave, Director of Product Marketing, with RiOS 5.0, Riverbed enables even faster performance for enterprise applications and enhances the manageability to deliver its market-leading WDS solution.
Riverbed accelerates the access to information to improve processes and make businesses more agile. With improved network flexibility, security and ease of management for large-scale deployments, it has extended the wide-area data services (WDS) solutions leadership, and helps data and businesses move faster. Significantly, Riverbed's growth rate outpaces the industry.
Davé, said: "CIOs realize this to be an absolute part of their investment. People have seen bandwidth savings in major applications." Citing globalization as a major challenge, he added that organizations needed to solve applications performance.
Network flexibility
Riverbed provides flexibility for organizations that want to enable end-to-end transparency of WAN traffic by IP address and port or by port only. Steelhead appliances enable full IP or port-only address visibility on the WAN for optimized connections in addition to their default correct addressing mode.
The security functionality in RiOS 5.0 builds on Riverbed’s extensive work in building solutions that ensure the security and integrity of enterprise data. With RiOS 5.0, Riverbed has enhanced the implementation of SSL acceleration, with new functionality for ease-of-setup and management. The QoS capability provides granularity, as well as simplified management for complex environments.
The company's Steelhead products accelerate applications to branch offices and mobile workers. Apparently, Steelhead products cut bandwidth usage by 60-95 percent.
Meeting top CIO priorities for 2008
CIOs are said to trying to design around four principles -- to support an agile enterprise. These are -- flexibility, simplicity, continuity and security. So what are the top priorities for CIOs in 2008? As per a December 2007 Goldman Sachs CIO Survey, the top 20 priorities for CIOs are:
1. Business intelligence; 2. Application integration; 3. ERP software; 4. BPM/workflow software; 5. Security; 6. Storage hardware (disk arrays, appliances, etc.); 7. DR/BCP; 8. Server virtualization; 9. Server consolidation; 10. Data center consolidation; 11. Web application development, including portals; 12. Wireless networking; 13. Mobile computing/remote access; 14. Compliance/risk management; 15. WAN optimization; 16. Storage management software (including backup); 17. SOA; 18. Cost cutting; 19. CRM software; 20. Thin client computing.
And where does WDS fit in with CIO priorities? In all of them, barring All of them, except nos. BI, BPM/workflow software, storage hardware and storage management software! WDS is based on accelerating the enterprise applications!
The Riverbed Optimization System takes care of data streamlining -- optimize WAN bandwidth utilization, prioritize applications' bandwidth use, and accelerate large-scale disaster recovery jobs; transport streamlining -- eliminate transport protocol inefficiencies, support secure traffic acceleration (SSL); application streamlining -- optimize application WAN performance and reduce protocol chatiness; and management streamlining -- enable transparent deployment, provide additional best-of-breed services, and facilitate central management and reporting.
The enterprise can eliminate bandwidth constraints. In fact, it can enjoy up to 5x more bandwidth without additional network costs. Applications can be accelerated over the WAN to mobile users or branch offices up to 100x times faster. Make it simple to deploy and manage 15 minutes for basic deployment enable consolidation. WDS is said to have a tangible, fast payback. An IDC study finds 7.3 month payback for Riverbed implementations.
Top features in RiOS 5.0
Riverbed's RiOS 5.0 delivers choice for enterprises through market-leading features. Two very significant features are accelerating Exchange and the RiOS Services Platform (RSP).
In RiOS 5.0, only one module may be run in the RSP, and not all services may be currently available. The RSP runs in a protected zone and does not impact the resource allocations dedicated to maintain peak Steelhead operations.
Dave elaborated: "We have created an area on our software where users can deploy best-of-breed services. In the first release, we will allow only one among print services, streaming media and networking (DNS/DHCP). In the next phase, we will enable a broad range of services." Enterprises can therefore leverage the WDS footprint further and also use best-of-breed services from other vendors.
Riverbed has also been very focused on Web based applications. He added: "We have introduced new features that are Web applications. In rel 5.0, we have added three applications/features."
Showing posts with label CIOs. Show all posts
Showing posts with label CIOs. Show all posts
Wednesday, April 30, 2008
Monday, December 17, 2007
10GbE Blade architectures suitable for next-gen data centers
What are some of the critical customer challenges today? For starters, server and storage sprawl increases operational complexity and reduces business predictability. Next, uncontrolled growth has resulted in lack of: space, power and CRAC capacity.
However, it seems that Blade server architectures can alleviate most of data center resource issues. In this context, if customers consolidate, can the network infrastructure optimally sustain the increased workload? And more importantly, what network technology would make sense for the environment?
I recently managed to speak with Adam Mendoza, Senior Manager, OS Alliances, NetApp, and Joel Reich, GM & Sr. Director, SAN/iSAN Business Unit, NetApp.
They attempted to answer these tricky customer pain points, and also touched upon how end-to-end 10GbE Blade architectures could lower IT infrastructure costs, prevent data center equipment sprawl, and help implement consolidation and virtualization.
For the record, recently, members of Blade.org -- BLADE Network Technologies Inc. (BLADE), a leading provider of network switching infrastructure for blade servers; Chelsio Communications (Chelsio), a leading provider of 10GbE network interface cards for servers and storage; and Network Appliance Inc. (NetApp), a leading provider of storage and data management solutions -- completed a collaborative study through Blade.org!
The study demonstrated that blade servers equipped with embedded 10 Gigabit Ethernet (10GbE) networking can support large-scale business applications at near wire-speeds in consolidated blade server architectures using NetApp Ethernet and IP networked storage.
The study, conducted on the IBM BladeCenter platform, found that by supporting multiple protocols, workloads and integrated data management across a 10GbE network, customers can protect their investment, lower IT infrastructure costs, prevent data center equipment sprawl, and implement consolidation and virtualization without sacrificing flexibility. Here's an image that shows this!
Virtualization a powerful trend
Even though virtualization has been doing the rounds for a couple of years now, it has not exactly reached where it should. On the uptake of virtualization by enterprises, Reich said: "What we've seen is probably for virtualization that VMWare is dominant. There's lot of hype around it.
"In reality, virtualization the most important and powerful trend. It is the real big wave of open system computing replacing maniframe applications. People have become smart and got very good at separating hype from real benefit of technology. Virtualization is not quite there as people are using it to consolidate servers, etc., that are running tier 1 applications."
Reich added that people were having big success in consolidating on virtualization in diverse tier 3 applications hosted on Wintel servers. Also, some tier 2 applications were getting consolidated as well. According to him, applications that don't drive critical revenue or manufacturing process would be the tier 3 applications. Tier 2 applications would include small data warehouses, etc.
He explained: "What's going on now is, lot of the introduction of virtualization was in tier 3 applications. It was opportunistic! Customers didn't have to put in lot of architechture thought into tier 3 and tier 2 apps, but they need to do in tier 1 apps. People also need to see what their data centers need to look like, especially when they move to virtualized environment. Silos are purpose-built spaces dedicated to an application in a data center. You need to have specific skill sets to manage those silos."
Blade.org's work
Commenting on Blade.org's work, Mendoza, said: "Effectively, when we first became involved with Blade, the intent was to collectively provide solutions end-to-end. NetApp's one of the few storage companies involved. We came up with a concept of characterizing what business critical applications form across multiple protocols. We need to have all different companies and groups collaborate and get the right experts involved. Blade.org gave us the environment and advantage."
As for the survey itself, they said: "We interviewed over 25 companies -- end users from Fortune 500 and channel partners, VARs, etc. We asked -- if we do this project, what are the things they would want to gain out of it?"
The findings were clear. First, this infrastructure could provide very good performance in this kind of consolidated environment. Second -- there has to be a comparison. "We are comparing networking components of 10GigE agianst fiber channel and utilizing different communications protocols," added Mendoza.
Scoring big with 10GbE
Over the last two yrs, 10GbE, customers have been saying that it is an interesting technology, but they didnt see applicability yet. There's also virtualization. So, if customers bring all this together, where will their bottlenecks be?
Adding the next layer of data management perspective, one of the main concerns of consolidation of this hype would be -- what about data protection, DR implementation, etc.? In this respect, NetApp brings the best kind of software that can handle all of those.
Reich added: "We have over 1,500 blade servers at our North Carolina facility. We learned how to manage that many servers. We take applications, data sets, etc., of blade servers, and save them on to the storage network. When concept customers or development teams say that they need something, we can provide that. We have already had the experience of large-scale dynamic environments. At the end of day, this is all applying to what customers want to see. We are collaborating to provide that to customers."
Another important result of the 10GbE has been an increase in the throughput. Mendoza said they provided the results of running these tests. "We did studies across iSCSI, fiber channel, etc. We are better than 2X in our implementation across 10GE."
So, what can be done to get servers consolidated? How can the data center equipment sprawl be reduced? Reich added: "In virtualization implementations, now, we are talking about a single housing that contains 14 servers. If you start virtualizaing over these 14 servers, you get one footprint. Space wise -- 14 servers come into a 3x2 space. In this manner, we reduced server sprawl, network sprawl and storage sprawl." This also has implications as far as power, cooling, etc., are concerned.
As for the 10GbE aspect, Reich said: "In many cases of a virtualized environment, its important to ask -- what's the new data center that I build will look like? People are also looking for information about how the network infrastructure should look like. Their reason: Ethernet is important as a storage interconnect, and that's because most data centers have two networks -- fiber channel or Ethernet."
So, if you can build something that's going to be ready in two years, can a data center be built that only has one type of network? Both, Reich and Mendoza accept that there's much more work to be do as users really need this kind of information.
"This is just the first step. It is such an important topic that there's a very great need for information," Reich said.
Mendoza added: "Our solution can simplify fabric topology in today's datacenters without compromising reliability? The big question that customers have is: why do we need multiple networks? Fabric refers to the network."
He said: "Part of the reason is simplicity and also economic. The way blade centers are built, you can add different connections to a blade center. The standard configuration of a blade center is the Ethernet. The economic considerations are: you can spend lot less money on blade center infrastructure. The Ethernet is the basic offering. That's part of the reason why, we focused on the 10GbE (Ethernet) from a cost standpoint."
Customers are said to be gaining confidence from the results of the solution development collaboration through Blade.org! Virtualization is the next killer application. It would allow a mixed nature of I/O and CPU utilization, add in standard data protection functions, top it off with application and guest OS management, etc. Customers have also indicated that they have deployed upward of 40 guest OS instances on a single server.
It must be noted that Blade server architectures require the same administrative processes: install, configure, customize, manage, optimize, etc. However, consolidation requires whole new approaches to where the data resides and the speed of access -- both in bandwidth and best practices. The 10GbE Blade architectures are the next-generation data center and are ready for business critical implementations.
However, it seems that Blade server architectures can alleviate most of data center resource issues. In this context, if customers consolidate, can the network infrastructure optimally sustain the increased workload? And more importantly, what network technology would make sense for the environment?
I recently managed to speak with Adam Mendoza, Senior Manager, OS Alliances, NetApp, and Joel Reich, GM & Sr. Director, SAN/iSAN Business Unit, NetApp.
They attempted to answer these tricky customer pain points, and also touched upon how end-to-end 10GbE Blade architectures could lower IT infrastructure costs, prevent data center equipment sprawl, and help implement consolidation and virtualization.
For the record, recently, members of Blade.org -- BLADE Network Technologies Inc. (BLADE), a leading provider of network switching infrastructure for blade servers; Chelsio Communications (Chelsio), a leading provider of 10GbE network interface cards for servers and storage; and Network Appliance Inc. (NetApp), a leading provider of storage and data management solutions -- completed a collaborative study through Blade.org!
The study demonstrated that blade servers equipped with embedded 10 Gigabit Ethernet (10GbE) networking can support large-scale business applications at near wire-speeds in consolidated blade server architectures using NetApp Ethernet and IP networked storage.
The study, conducted on the IBM BladeCenter platform, found that by supporting multiple protocols, workloads and integrated data management across a 10GbE network, customers can protect their investment, lower IT infrastructure costs, prevent data center equipment sprawl, and implement consolidation and virtualization without sacrificing flexibility. Here's an image that shows this!
Virtualization a powerful trendEven though virtualization has been doing the rounds for a couple of years now, it has not exactly reached where it should. On the uptake of virtualization by enterprises, Reich said: "What we've seen is probably for virtualization that VMWare is dominant. There's lot of hype around it.
"In reality, virtualization the most important and powerful trend. It is the real big wave of open system computing replacing maniframe applications. People have become smart and got very good at separating hype from real benefit of technology. Virtualization is not quite there as people are using it to consolidate servers, etc., that are running tier 1 applications."
Reich added that people were having big success in consolidating on virtualization in diverse tier 3 applications hosted on Wintel servers. Also, some tier 2 applications were getting consolidated as well. According to him, applications that don't drive critical revenue or manufacturing process would be the tier 3 applications. Tier 2 applications would include small data warehouses, etc.
He explained: "What's going on now is, lot of the introduction of virtualization was in tier 3 applications. It was opportunistic! Customers didn't have to put in lot of architechture thought into tier 3 and tier 2 apps, but they need to do in tier 1 apps. People also need to see what their data centers need to look like, especially when they move to virtualized environment. Silos are purpose-built spaces dedicated to an application in a data center. You need to have specific skill sets to manage those silos."
Blade.org's work
Commenting on Blade.org's work, Mendoza, said: "Effectively, when we first became involved with Blade, the intent was to collectively provide solutions end-to-end. NetApp's one of the few storage companies involved. We came up with a concept of characterizing what business critical applications form across multiple protocols. We need to have all different companies and groups collaborate and get the right experts involved. Blade.org gave us the environment and advantage."
As for the survey itself, they said: "We interviewed over 25 companies -- end users from Fortune 500 and channel partners, VARs, etc. We asked -- if we do this project, what are the things they would want to gain out of it?"
The findings were clear. First, this infrastructure could provide very good performance in this kind of consolidated environment. Second -- there has to be a comparison. "We are comparing networking components of 10GigE agianst fiber channel and utilizing different communications protocols," added Mendoza.
Scoring big with 10GbE
Over the last two yrs, 10GbE, customers have been saying that it is an interesting technology, but they didnt see applicability yet. There's also virtualization. So, if customers bring all this together, where will their bottlenecks be?
Adding the next layer of data management perspective, one of the main concerns of consolidation of this hype would be -- what about data protection, DR implementation, etc.? In this respect, NetApp brings the best kind of software that can handle all of those.
Reich added: "We have over 1,500 blade servers at our North Carolina facility. We learned how to manage that many servers. We take applications, data sets, etc., of blade servers, and save them on to the storage network. When concept customers or development teams say that they need something, we can provide that. We have already had the experience of large-scale dynamic environments. At the end of day, this is all applying to what customers want to see. We are collaborating to provide that to customers."
Another important result of the 10GbE has been an increase in the throughput. Mendoza said they provided the results of running these tests. "We did studies across iSCSI, fiber channel, etc. We are better than 2X in our implementation across 10GE."
So, what can be done to get servers consolidated? How can the data center equipment sprawl be reduced? Reich added: "In virtualization implementations, now, we are talking about a single housing that contains 14 servers. If you start virtualizaing over these 14 servers, you get one footprint. Space wise -- 14 servers come into a 3x2 space. In this manner, we reduced server sprawl, network sprawl and storage sprawl." This also has implications as far as power, cooling, etc., are concerned.
As for the 10GbE aspect, Reich said: "In many cases of a virtualized environment, its important to ask -- what's the new data center that I build will look like? People are also looking for information about how the network infrastructure should look like. Their reason: Ethernet is important as a storage interconnect, and that's because most data centers have two networks -- fiber channel or Ethernet."
So, if you can build something that's going to be ready in two years, can a data center be built that only has one type of network? Both, Reich and Mendoza accept that there's much more work to be do as users really need this kind of information.
"This is just the first step. It is such an important topic that there's a very great need for information," Reich said.
Mendoza added: "Our solution can simplify fabric topology in today's datacenters without compromising reliability? The big question that customers have is: why do we need multiple networks? Fabric refers to the network."
He said: "Part of the reason is simplicity and also economic. The way blade centers are built, you can add different connections to a blade center. The standard configuration of a blade center is the Ethernet. The economic considerations are: you can spend lot less money on blade center infrastructure. The Ethernet is the basic offering. That's part of the reason why, we focused on the 10GbE (Ethernet) from a cost standpoint."
Customers are said to be gaining confidence from the results of the solution development collaboration through Blade.org! Virtualization is the next killer application. It would allow a mixed nature of I/O and CPU utilization, add in standard data protection functions, top it off with application and guest OS management, etc. Customers have also indicated that they have deployed upward of 40 guest OS instances on a single server.
It must be noted that Blade server architectures require the same administrative processes: install, configure, customize, manage, optimize, etc. However, consolidation requires whole new approaches to where the data resides and the speed of access -- both in bandwidth and best practices. The 10GbE Blade architectures are the next-generation data center and are ready for business critical implementations.
Subscribe to:
Posts (Atom)
