My next week will be very busy because ITEXPO and other industry events are running in parallel in Miami, Florida. I looked through the schedule and - as of today - I will be presenting in five sessions: three of them are part of ITEXPO while the other two are part of the Ingate Summit.
My first session at ITEXPO is on February 2 at 1:30pm local time and is titled “How Does The Traditional Desktop Phone Fit Into The Evolving Enterprise User Experience?” Frank Stinson from Intellicom Analytics is moderating and I will be sitting next to speakers representing other business phone makers. This session will explore how the trend to Unified Communications impacts the way people access voice services – through telephones but also through soft clients, smart phones, and tablet devices. How should desktop telephones evolve in this environment and how are desktop phone manufacturers planning to increase the value of their products given evolving user expectations?
Then on February 3 at 1pm I will present in the session “Making Telepresence Affordable and Reliable” which will be moderated by TMC Executive Director Paula Bernier and will also include Matt Collier from LifeSize Communications. This session will discuss the perception that telepresence is expensive and will clog your network with HD video traffic. My talk will focus on the reduced network bandwidth consumption (i.e. minimizing or avoiding IP network upgrades) and on the new network architectures that allow for virtualization of conference resources shared within the entire organization or deployed by service providers to serve vast user communities.
My last ITEXPO presentation is in the session “UC Interoperability” on February 3 at 2pm. David Yedwab from Partner Market Strategies and Analytics will moderate and there will be two other panelists: Alan Percy from AudioCodes and Allen Mendelshon from Avaya’s UC Strategy team. The session will address the need for interop and standards in multivendor environments and explore different aspects of multi-vendor UC interoperability.
The Ingate Summit is organized in parallel to ITEXPO and I had the pleasure to present at previous Summits, the last one being in Los Angeles in October. This time, the Summit starts early with pre-conference service provider workshops and I will present in the session "Generating Revenue from HD Video” on February 1 at 5:30pm. Joel Maloff of Maloff NetResults is moderating, and I will share the time with Karl Stahl from Intertex Data. Since the audience is serive providers, I will focus on the managed and hosted telepresence services, and also address ITSPs with current hosted voice offering that would like to add HD video services without much CAPEX. I will also provide an update of the industry efforts in the areas of telepresence interoperability and B2B video communications.
My second session at the Ingate Summit is the "Town Hall Meeting: Unified Communications" on February 3 at 9am. The list of panelists is fairly long: Chad Krantz from Brodvox, Dan York from VOIPSA, Karl Stahl from Intertex, Jeff Ridley from ShoreTel, David Yedwab from Market Strategy & Analytics Partners, and Gary Mading from Aastra. I will be wearing my UCIF hat in that session, i.e., representing the Unified Communications Interoperability Forum. In such large panel, there will be no time for slides but we will have a discussion around the scope of Unified Communications, and how different vendors approach UC. I will focus on the UCIF philosophy (certification, not standards development) and call for other companies to join and influence the discussions in UCIF.
All in all, next week in Miami will be very hot - although they do tend to set the air conditioning in the Miami Beach Convention Center on “very cold”. I am sure there will be a lot of interesting discussions and I hope to meet some of the blog readers there in person. For the rest, I would recommend watching the video interview that I will give on the last day of the conference (February 4). I will summarize the news from the conference in my answers during this interview. Once the link to the recording becomes available, I will add it to this blog post.
Friday, January 28, 2011
Monday, January 24, 2011
Industry events, speaking engagements, and white papers
In addition to blog posts, Video Networker keeps a complete list of the industry events that I attend and the topics of my speaking engagements. It also has a section with links to my white papers. See below!
Monday, December 20, 2010
The Hottest Topics in 2010
Year 2010 is coming to an end and this gives me a great opportunity to look back, and summarize the major communications market developments in the past 12 months. Unified Communications was the most important topic in 2010, and was frequently discussed at industry events (see links below), in online forums, and in papers.
While most of the technology discussions were around better ways to compress video – for example, via H.264 High Profile and Scalable Video Coding – business discussions focused mainly on cloud-based UC services that service providers are starting to explore. Unresolved interoperability issues in the area of Telepresence and Unified Communications were discussed at meetings of IMTC, UCIF, and standardization organizations while HD Voice continued to penetrate new market segments.
New Video Technologies
One of the hottest new technology developments in 2010 was the implementation of H.264 High Profile which allows for substantial network bandwidth savings when transmitting HD or SD video over the network. To describe the technology and its benefits, I wrote the white paper “H.264 High Profile: The Next Big Thing in Visual Communications” in April, and since the technology was moving so fast, I had to update the paper in June.
By October there was so much interest around High Profile in the service provider community that my colleague Ian Irvine and I put together an additional paper “The Opportunity for Service Providers to Grow Business with Polycom High Profile” which looks at the High Profile technology from a service provider perspective.
Another hot technology topic in 2010 was H.264 Scalable Video Coding, and in November I wrote the paper “More Scale at Lower Cost with Scalable Video Coding” that discusses the SVC benefits but also the interoperability challenges surrounding this new technology.
Cloud-based UC Services
While VOIP service providers have been around for a while, UC services provided by service providers are quite new phenomenon. In 2010, I had the pleasure to meet several service providers – Simple Signal is one example – that have created amazing services using UC technology.
The Broadband World Forum in October was great opportunity to invite service providers and UC vendors for a panel discussion around UC. I chaired the session “Unified Communication in the Cloud” which examined which UC functions bring real value to enterprises and how hosted or managed service providers can help deliver this UC functionality. Andreas Arrigoni, Head of Collaboration Services at Swisscom, and David Gilbert, President of SimpleSignal, provided perspectives on the service provider role for rolling out UC services.
The UC vendor community was represented by Thierry Batut, Sales Leader for Unified Communications at IBM Europe, George Paolini, VP Business Solutions at Avaya, Glynn Jones, VP at Polycom EMEA, and David Noguer Bau, Head of SP Multiplay Marketing EMEA at Juniper Networks. The 90-minute session was an excellent mix of presentations and panel discussion, and is a model I would love to replicate at other industry events in the future.
Cloud services are definitely the new frontier for Unified Communication, and I have summarized my thought about voice and video collaboration services in the cloud in a blog post in September. Then in October BroadSoft announced its new BroadCloud service, and I spent some time describing how Polycom and BroadSoft partnered over the years – in fact, since 2002 – to first make hosted VOIP robust and easy to deploy, and now expand to video services in the BroadCloud. Read the entire story in the new white paper “From Hosted Voice to Cloud-Based Unified Communication Services”.
Interoperability
In my blog post in May, I made the argument for standards and interoperability, and for the need of additional test and certification work in the Unified Communications Interoperability Forum (UCIF). Then I had the chance to present on “How UC and Telepresence Are Changing Video Protocols and Interoperability Forever” at the Wainhouse Research Collaboration Summit in June. Interoperability – and more specifically telepresence Interoperability - was the most important topic at IMTC SuperOp!, which was followed by a lot of face-to-face and virtual discussions throughout 2010. The most recent one was a roundtable with telepresence vendors organized by Howard Lichtman from the Human Productivity Lab.
UC interoperability work started in the UCIF and the forum’s first face-to-face meeting took place in October. The true value UCIF brings to the table is the ability to create test specifications, verify, and certify vendor compliance and interoperability. This will finally create an independent seal of approval that is very much missing in UC environments today, and which customers are calling for before committing to Unified Communications.
HD Voice Everywhere
In 2010, HD Voice finally moved from on-premise installations in the enterprise to scalable services provided by tier-one service providers. The third HD Communications Summit in February was the first European HD Voice event and was hosted by Orange. The Summit gathered voice industry professionals from across Europe and the United States to discuss the state of HD Voice deployments and future plans. It was a great opportunity to meet some old friends in the voice industry and some new people, to check deployment progress, and compare notes on where the voice industry is going.
Happy Holidays
I would like to thank all Video Networker followers for their continuous support and great feedback in 2010. Very best wishes to you and your families in this holiday season!
While most of the technology discussions were around better ways to compress video – for example, via H.264 High Profile and Scalable Video Coding – business discussions focused mainly on cloud-based UC services that service providers are starting to explore. Unresolved interoperability issues in the area of Telepresence and Unified Communications were discussed at meetings of IMTC, UCIF, and standardization organizations while HD Voice continued to penetrate new market segments.
New Video Technologies
One of the hottest new technology developments in 2010 was the implementation of H.264 High Profile which allows for substantial network bandwidth savings when transmitting HD or SD video over the network. To describe the technology and its benefits, I wrote the white paper “H.264 High Profile: The Next Big Thing in Visual Communications” in April, and since the technology was moving so fast, I had to update the paper in June.
By October there was so much interest around High Profile in the service provider community that my colleague Ian Irvine and I put together an additional paper “The Opportunity for Service Providers to Grow Business with Polycom High Profile” which looks at the High Profile technology from a service provider perspective.
Another hot technology topic in 2010 was H.264 Scalable Video Coding, and in November I wrote the paper “More Scale at Lower Cost with Scalable Video Coding” that discusses the SVC benefits but also the interoperability challenges surrounding this new technology.
Cloud-based UC Services
While VOIP service providers have been around for a while, UC services provided by service providers are quite new phenomenon. In 2010, I had the pleasure to meet several service providers – Simple Signal is one example – that have created amazing services using UC technology.
The Broadband World Forum in October was great opportunity to invite service providers and UC vendors for a panel discussion around UC. I chaired the session “Unified Communication in the Cloud” which examined which UC functions bring real value to enterprises and how hosted or managed service providers can help deliver this UC functionality. Andreas Arrigoni, Head of Collaboration Services at Swisscom, and David Gilbert, President of SimpleSignal, provided perspectives on the service provider role for rolling out UC services.
The UC vendor community was represented by Thierry Batut, Sales Leader for Unified Communications at IBM Europe, George Paolini, VP Business Solutions at Avaya, Glynn Jones, VP at Polycom EMEA, and David Noguer Bau, Head of SP Multiplay Marketing EMEA at Juniper Networks. The 90-minute session was an excellent mix of presentations and panel discussion, and is a model I would love to replicate at other industry events in the future.
Cloud services are definitely the new frontier for Unified Communication, and I have summarized my thought about voice and video collaboration services in the cloud in a blog post in September. Then in October BroadSoft announced its new BroadCloud service, and I spent some time describing how Polycom and BroadSoft partnered over the years – in fact, since 2002 – to first make hosted VOIP robust and easy to deploy, and now expand to video services in the BroadCloud. Read the entire story in the new white paper “From Hosted Voice to Cloud-Based Unified Communication Services”.
Interoperability
In my blog post in May, I made the argument for standards and interoperability, and for the need of additional test and certification work in the Unified Communications Interoperability Forum (UCIF). Then I had the chance to present on “How UC and Telepresence Are Changing Video Protocols and Interoperability Forever” at the Wainhouse Research Collaboration Summit in June. Interoperability – and more specifically telepresence Interoperability - was the most important topic at IMTC SuperOp!, which was followed by a lot of face-to-face and virtual discussions throughout 2010. The most recent one was a roundtable with telepresence vendors organized by Howard Lichtman from the Human Productivity Lab.
UC interoperability work started in the UCIF and the forum’s first face-to-face meeting took place in October. The true value UCIF brings to the table is the ability to create test specifications, verify, and certify vendor compliance and interoperability. This will finally create an independent seal of approval that is very much missing in UC environments today, and which customers are calling for before committing to Unified Communications.
HD Voice Everywhere
In 2010, HD Voice finally moved from on-premise installations in the enterprise to scalable services provided by tier-one service providers. The third HD Communications Summit in February was the first European HD Voice event and was hosted by Orange. The Summit gathered voice industry professionals from across Europe and the United States to discuss the state of HD Voice deployments and future plans. It was a great opportunity to meet some old friends in the voice industry and some new people, to check deployment progress, and compare notes on where the voice industry is going.
Happy Holidays
I would like to thank all Video Networker followers for their continuous support and great feedback in 2010. Very best wishes to you and your families in this holiday season!
Thursday, October 14, 2010
Inside the Unified Communications Interoperability Forum (UCIF)
First UCIF Face-to-Face Meeting
UCIF members met for the first time face-to-face during IT EXPO in Los Angeles last week. I had four presentations at IT EXPO and was in town, so I had the opportunity to meet key people and get an overview of the activities in UCIF. Polycom is a founding member of the forum, and actively participates in working groups along with Microsoft, Logitech/LifeSize, and other member companies.
36 people gathered in the LA Convention Center for series of meetings on Tuesday and Wednesday, while up to 22 people joined online. Recruitment of additional members is ongoing, so if your company would like to join, let me know.
UCIF Working Groups
The meeting included sessions of the active UCIF groups. Each UCIF group starts as a Study Group (similar to a BOF in IETF) which studies certain issue, and develops a proposal for charter. Once the charter is approved by the UCIF board, the group becomes a Task Group. On a high level, UCIF has three WGs: Technical WG, Test & Certification WG, and Marketing WG. Within the Technical WG, there are currently three Task Groups (USB Audio Task Group, Webcam Task Group, and H.264 Profile Task Group), and three Study Groups (Voice Study Group, Instant Messaging and Presence Study Group, and Provisioning Study Group).
I was able to attend few of the groups’ sessions, and got the following understanding of where UCIF is going. Based on the trend towards simplifying the infrastructure for video (and preparing it for cloud deployments), Scalable Video Coding is the focus of the H.264 Profile Task Group. The group will create an H.264 SVC profile to ensure that video encoders and decoders interoperate. The group will not address SVC description in SDP, transport via RTP, etc., since there is already work on these topics at IETF.
SVC moves the complexity from the infrastructure to the edges of the network, i.e. in the endpoints. As video soft client proliferation is expected to lead to mass deployment of video, solutions are needed to free the computer CPU from the video processing work. Current SVC implementations run on high-end computers and consume a lot of their performance impacting other applications; therefore, looking for ways to free the CPU is an honorable task.
The UCIF Webcam Task Group is focusing on compressing video in the webcam, and defining the interface between webcam and video soft client application. Since webcams are usually connected to the computer via USB, the USB Video Class Specification V1.1 can be used as baseline but H.264 SVC configuration requires exchange of additional parameters. The specification must also allow for multiplexing video streams on single USB interface while keeping the interface simple for the client.
Provisioning of endpoints is critical in multi-vendor environments where the configuration server may come from one vendor while the endpoint comes from another. Most recently, the SIP Forum set out to define a profile and recommendations for User Agent configuration. The result of this work was a contribution in IETF that describes a mechanism for server discovery using existing standards. The SIP Forum, however, did not address the configuration data model/schema. At its meeting, the UCIF Provisioning Study Group discussed the gap between its charter and the work done in the SIP Forum. There does not seem to be a good reason for UCIF to define yet another mechanism for discovery for SIP User Agents when both SIP Forum and IETF have defined mechanisms. Focusing on the format for the configuration data (data model/schema) makes sense and so does creating a test and certification around the provisioning interface. Several previous attempts at IETF to define a schema can be used as a starting point for the UCIF Provisioning SG.
The UCIF Provisioning Group is still a Study Group but a charter is almost done and nothing should stay in the way for the group to become Task Group very soon. Polycom is obviously very interested in this work, since Polycom endpoints – telephones, video endpoints, etc. – are deployed with dozens of servers. Wouldn’t it be nice to have a standard way of configuring endpoints, no matter what environment the endpoints are deployed in? Another example for the importance of interoperability is deployment of CMA 5000 management application in mixed video endpoint environments. Now management applications have to support multiple configuration methods to support endpoints from Polycom, Tandberg, etc., which is very inefficient and limits scale. The UCIF Provisioning group discussed the use of service announcement via DNS SRV and Bonjour (previously called ‘Rendezvous’). While discovering the provisioning service is important, I think defining the configuration format - possibly an XML file – and transport mechanism – I vote for HTTPS – are important requirements for interoperability.
The Test and Certification Work Group is probably the most important group of all. All working groups are required to create test plans along with creating interop specifications. When a group finishes work, it sends the interop specification and the test plan to the Test and Certification Group, which is responsible for testing and certification of the vendors that pass the test. This promises more structured approach to testing than other test venues such as SuperOp and SIPit.
UCIF and Other Industry Organizations
I had the chance to represent UCIF in several industry panels, and the question that comes up a lot is how UCIF relates to other industry forums, for example, the International Multimedia Telecommunications Consortium (IMTC) and the SIP Forum.
Marc Robins, President of the SIP Forum, and Richard Shockey, chair of SIP Forum’s board, attended the UCIF meeting in L.A., and identified areas of possible cooperation. Founded as a nonprofit in 2000, the SIP Forum includes many of the same companies that participate in UCIF, and focuses on SIP Trunking (based on the SIP Connect specification, now in V1.1), User Agent Configuration, and Fax-Over-IP. UCIF is currently not looking at fax, although surprisingly fax is getting a second wind with mandatory HIPAA requirements that do not allow sending medical test results over email. The work in the SIP Forum’s User Agent Configuration group led to the definition of a mechanism for finding configuration servers but stops short of defining the actual format of configuration files. The UCIF Provisioning Group could take that work to the next level, define formats, create test plans, and certify vendors that comply. The same applies to the SIP trunking work in the SIP Forum: the UCIF Voice Study Group could take the SIP Connect specification, create a test plan, and certify vendors complying with it. I think that if the group focuses on SIP trunking (or more generally on SIP interop), it should not be called Voice Group. While SIP trunks today are defined and used for voice, they will support video once SPs start interconnecting video IP-PBXs through SIP trunks.
With regards to IMTC, I see interest in telepresence interoperability in both IMTC and UCIF. Standardization bodies ITU-T and IETF are also working on the subject. Whatever happens in this area, I expect that any activities in UCIF will include test plans and certification, which is not in the scope for IMTC and standards organizations. Obviously, the challenge with interop test among telepresence systems is the size of these systems, and the difficulties moving them. This leads to the requirement for testing infrastructure to connect UCIF members over the Internet in a test environment that allows for continuous testing.
Summary
The true value UCIF brings to the table is the ability to create test specifications, verify, and certify vendor compliance and interoperability. This will finally create an independent seal of approval that is very much missing in UC environments today, and which customers are calling for before committing to Unified Communications.
Wednesday, September 1, 2010
Voice and Video Collaboration Services in the Cloud
Cloud computing is defined as ‘virtualization of computing assets delivered on demand over the IP network’. It promises availability and scalability for applications ranging from storage to collaboration. Clouds are in particularly popular because the concept is easier to grasp than previous attempts to define similar services through Application Service Providers (ASPs) and Software as a Service (SaaS).
The Cloud is a more general concept that includes not only SaaS but also storage, platform, and infrastructure as a service. Clouds are in better position to deliver on the promise for interactivity. While slow networks in the past have made the user experience with ASPs quite negative, better networks available today allow for fast response times, and increased interactivity.
Analysts are excited about cloud services and see above-average growth; while the average IT market growth is expected to be 4% per year until 2013, IT cloud services are expected to grow 25% over same period. The current bidding war between HP and Dell for cloud storage technology provider 3PAR is a great example for how hot this market segment has become.
Evolution of Service Architectures
In the legacy approach, each enterprise application runs on separate server(s) that resides in one of the enterprise’s offices. This leads to inefficient use of space, energy (power-up and cooling), and substandard user experience. In the next stage of the evolution, all servers were collocated in a data center where they can share space and power. In the third stage, services are provided by the Cloud that can be within the enterprise (‘private cloud’) or outside the enterprise (‘public cloud’).
The term ‘virtualization’ is often used in relation to Clouds, and it is important to clarify virtualization’s role. Virtualization saves money by increasing server utilization, i.e. reducing the number of servers (hardware) necessary to support applications. Virtualization can be used in a traditional data center or in the Cloud. In both environments, virtualization reduces the hardware necessary to run enterprise applications. It has very strong positive financial and environmental impact.
Unified Collaboration
Unified collaboration combines variety of communication tolls – voice, video, email, presence, IM, etc. – into a seamless user experience, and into workflows, through a single user interface. Since UC installations are far past pilot deployments and are now being rolled out across large organizations, scalability is an important requirement.
Global teams span the entire world, and different time zones do not allow everyone on the team to participate live in all collaboration sessions. Audio, video, and shared content must therefore be stored and streamed. This leads to the requirement for efficient and scalable storage.
Accessibility of UC applications has two sides. First, users should be able to access them from anywhere, not just the office bit also from remote locations. Second, any device should provide access, including computers, telephones, appliances such as personal and group video systems, and even immersive telepresence systems.
To meet these UC requirements, UC architectures must follow IT architectures towards Clouds.
Bandwidth Requirements
UC applications, such as voice and video, require higher quality of service (QOS) than applications such as email, scheduling, or management. QOS are defined by bandwidth, latency, packet loss, and jitter. And while there are mechanisms in place to combat packet loss and jitter, bandwidth remains the most important resource necessary to support voice and video collaboration applications.
If multiple systems have to be connected in a multi-point conference, the traffic quickly grows, and may overwhelm the Cloud. Cloud throughput is critical for successful deployment of voice and video collaboration application. Recent advances in video compression technology, in particular Polycom’s implementation of the H.264 High Profile for real-time video, allow for ‘thinner’ connections between premise and Cloud without sacrificing the quality of experience.
In general, voice transmission requires less bandwidth than video. Even the highest audio quality does not require more than 128 kbps per channel and this is usually not an issue for the interface between customer premise and the cloud service provider.
Security Requirements
Numerous surveys of CIOs and IT administrators have shown that security is the leading concern around deploying Cloud services. With data applications, hackers try to capture and copy the customer data. With real-time collaboration applications, such as voice and video, hackers try to redirect and record voice and video calls.
There is currently a fairly robust security framework for user authentication, authorization, and media encryption – both in SIP and H.323 environments - that can be deployed to prevent interception of voice and video calls at the interface between customer and SP. However, this security framework has to be reevaluated and extended to cover new security threats that come with new cloud service use cases.
Many industry experts believe that cloud services will lead to improved security due to the centralization of data and the increased security-focused resource. SPs are able to devote resources to solving security issues that many customers cannot afford to solve themselves.
Availability Requirements
In the cloud services scenario, a lot of the infrastructure functionality that today resides on customer premise will be moved to the Cloud, and become a shared resource among enterprises. The availability of these resources is of paramount importance to the success of voice and video services in the Cloud.
One successfully deployed approach to increased availability (and scale) is the use of a redundant resource management application that controls a pool of multipoint conferencing resources in the network.
To make a pool of conference servers behave as one huge conference server, the resource management application tracks incoming calls, routes them to the appropriate resource (for instance, this can be done based on available server resources but also based on available bandwidth to the location of this server) and that automatically creates cascading links if a conference overflows to another server. If the conference is prescheduled, the application server can select a conference server that has sufficient resources to handle the number of participants at the required video quality (bandwidth). Overflow situations are probable with ad-hoc conferences where participants spontaneously join without any upfront reservation of resources.
The resource management application runs on two servers to ensure 100% redundancy and auto-failover. It is designed to provide uninterrupted service by routing calls around failed or busy conference servers. It also allows administrators to “busy out” media severs for maintenance activities while still providing an ‘always available’ experience from the Cloud user point of view. The system can gradually grow from small deployments of 1-2 conference servers to large deployments with many geographically dispersed conference servers based on the dynamic needs of growing organizations. System administrators can monitor daily usage and plan the expansion as necessary. This approach also provides a centralized mechanism to deploy a front-end application to control and monitor conferencing activities across all conference servers.
The resource management application also serves as a load balancer in this scenario, that is, it distributes the conference load over a group of servers, ensuring that a server is not oversubscribed, while another being underutilized. The larger the resource pool, the more efficient the load balancing function is, a feature that is very important to Cloud service providers who can offer conference services globally by using the resource management application and placing conference servers in central points of their networks. More approaches to increased availability and scale are discussed in the paper ‘Polycom UC Intelligent Core: Scalable Infrastructure for Distributed Video’.
Bringing Collaboration and Clouds Closer
The trend towards cloud services is driving both technology and business model changes.
On the technology side, UC technology providers have to make significant changes in the architecture for voice and video applications to better align with the architecture of Clouds. Reducing the complexity in the infrastructure and pushing it to the endpoints is a viable approach although the impact on the user experience through complex endpoint implementation is still being evaluated.
Cloud service providers have to meet challenges on their own. Clouds today are designed with data processing in mind, and throughput (bandwidth to and from the Cloud) and Quality of Service (latency, for example) are not at the level required for real-time interaction. Cloud providers therefore will need to increase throughput for real-time applications, and develop new service pricing to accommodate the specifics of real-time collaboration.
The Cloud is a more general concept that includes not only SaaS but also storage, platform, and infrastructure as a service. Clouds are in better position to deliver on the promise for interactivity. While slow networks in the past have made the user experience with ASPs quite negative, better networks available today allow for fast response times, and increased interactivity.
Analysts are excited about cloud services and see above-average growth; while the average IT market growth is expected to be 4% per year until 2013, IT cloud services are expected to grow 25% over same period. The current bidding war between HP and Dell for cloud storage technology provider 3PAR is a great example for how hot this market segment has become.
Evolution of Service Architectures
In the legacy approach, each enterprise application runs on separate server(s) that resides in one of the enterprise’s offices. This leads to inefficient use of space, energy (power-up and cooling), and substandard user experience. In the next stage of the evolution, all servers were collocated in a data center where they can share space and power. In the third stage, services are provided by the Cloud that can be within the enterprise (‘private cloud’) or outside the enterprise (‘public cloud’).
The term ‘virtualization’ is often used in relation to Clouds, and it is important to clarify virtualization’s role. Virtualization saves money by increasing server utilization, i.e. reducing the number of servers (hardware) necessary to support applications. Virtualization can be used in a traditional data center or in the Cloud. In both environments, virtualization reduces the hardware necessary to run enterprise applications. It has very strong positive financial and environmental impact.
Unified Collaboration
Unified collaboration combines variety of communication tolls – voice, video, email, presence, IM, etc. – into a seamless user experience, and into workflows, through a single user interface. Since UC installations are far past pilot deployments and are now being rolled out across large organizations, scalability is an important requirement.
Global teams span the entire world, and different time zones do not allow everyone on the team to participate live in all collaboration sessions. Audio, video, and shared content must therefore be stored and streamed. This leads to the requirement for efficient and scalable storage.
Accessibility of UC applications has two sides. First, users should be able to access them from anywhere, not just the office bit also from remote locations. Second, any device should provide access, including computers, telephones, appliances such as personal and group video systems, and even immersive telepresence systems.
To meet these UC requirements, UC architectures must follow IT architectures towards Clouds.
Bandwidth Requirements
UC applications, such as voice and video, require higher quality of service (QOS) than applications such as email, scheduling, or management. QOS are defined by bandwidth, latency, packet loss, and jitter. And while there are mechanisms in place to combat packet loss and jitter, bandwidth remains the most important resource necessary to support voice and video collaboration applications.
If multiple systems have to be connected in a multi-point conference, the traffic quickly grows, and may overwhelm the Cloud. Cloud throughput is critical for successful deployment of voice and video collaboration application. Recent advances in video compression technology, in particular Polycom’s implementation of the H.264 High Profile for real-time video, allow for ‘thinner’ connections between premise and Cloud without sacrificing the quality of experience.
In general, voice transmission requires less bandwidth than video. Even the highest audio quality does not require more than 128 kbps per channel and this is usually not an issue for the interface between customer premise and the cloud service provider.
Security Requirements
Numerous surveys of CIOs and IT administrators have shown that security is the leading concern around deploying Cloud services. With data applications, hackers try to capture and copy the customer data. With real-time collaboration applications, such as voice and video, hackers try to redirect and record voice and video calls.
There is currently a fairly robust security framework for user authentication, authorization, and media encryption – both in SIP and H.323 environments - that can be deployed to prevent interception of voice and video calls at the interface between customer and SP. However, this security framework has to be reevaluated and extended to cover new security threats that come with new cloud service use cases.
Many industry experts believe that cloud services will lead to improved security due to the centralization of data and the increased security-focused resource. SPs are able to devote resources to solving security issues that many customers cannot afford to solve themselves.
Availability Requirements
In the cloud services scenario, a lot of the infrastructure functionality that today resides on customer premise will be moved to the Cloud, and become a shared resource among enterprises. The availability of these resources is of paramount importance to the success of voice and video services in the Cloud.
One successfully deployed approach to increased availability (and scale) is the use of a redundant resource management application that controls a pool of multipoint conferencing resources in the network.
To make a pool of conference servers behave as one huge conference server, the resource management application tracks incoming calls, routes them to the appropriate resource (for instance, this can be done based on available server resources but also based on available bandwidth to the location of this server) and that automatically creates cascading links if a conference overflows to another server. If the conference is prescheduled, the application server can select a conference server that has sufficient resources to handle the number of participants at the required video quality (bandwidth). Overflow situations are probable with ad-hoc conferences where participants spontaneously join without any upfront reservation of resources.
The resource management application runs on two servers to ensure 100% redundancy and auto-failover. It is designed to provide uninterrupted service by routing calls around failed or busy conference servers. It also allows administrators to “busy out” media severs for maintenance activities while still providing an ‘always available’ experience from the Cloud user point of view. The system can gradually grow from small deployments of 1-2 conference servers to large deployments with many geographically dispersed conference servers based on the dynamic needs of growing organizations. System administrators can monitor daily usage and plan the expansion as necessary. This approach also provides a centralized mechanism to deploy a front-end application to control and monitor conferencing activities across all conference servers.
The resource management application also serves as a load balancer in this scenario, that is, it distributes the conference load over a group of servers, ensuring that a server is not oversubscribed, while another being underutilized. The larger the resource pool, the more efficient the load balancing function is, a feature that is very important to Cloud service providers who can offer conference services globally by using the resource management application and placing conference servers in central points of their networks. More approaches to increased availability and scale are discussed in the paper ‘Polycom UC Intelligent Core: Scalable Infrastructure for Distributed Video’.
Bringing Collaboration and Clouds Closer
The trend towards cloud services is driving both technology and business model changes.
On the technology side, UC technology providers have to make significant changes in the architecture for voice and video applications to better align with the architecture of Clouds. Reducing the complexity in the infrastructure and pushing it to the endpoints is a viable approach although the impact on the user experience through complex endpoint implementation is still being evaluated.
Cloud service providers have to meet challenges on their own. Clouds today are designed with data processing in mind, and throughput (bandwidth to and from the Cloud) and Quality of Service (latency, for example) are not at the level required for real-time interaction. Cloud providers therefore will need to increase throughput for real-time applications, and develop new service pricing to accommodate the specifics of real-time collaboration.
Subscribe to:
Posts (Atom)