Pexip Infinity v9.1 Release Notes

Transcription

1 Pexip Infinity v9.1 Release Notes Contents Introduction 1 Upgrading to version New features and improvements in this release 3 Issues fixed in v9.1 6 Issues fixed in v9 7 Known limitations 9 Introduction This document contains the release notes for Pexip Infinity version 9.1, released on 16 April In addition, the following guides are available from the Pexip website ( They provide guidance on the installation, initial configuration, and operation of Pexip Infinity: Administrator Guide Getting Started Guide Server Design Guide Specifications and Requirements The Pexip website also contains comprehensive documentation on all aspects of deploying the Pexip Infinity platform. This includes how to use the Infinity Connect client suite, and how to integrate Pexip Infinity with other third-party systems and call control solutions including Microsoft Lync, Cisco Unified Communications Manager, Cisco VCS and Polycom DMA. Management Node host server sizing information You must increase the size of the Management Node host server to 2 cores and 4 GB of RAM for any deployments with more than 10 Conferencing Nodes. We recommend increasing to 4 cores and 6 GB of RAM for any deployments with more than 30 Conferencing Nodes Pexip AS Version 9.1.a May 2015 Page 1 of 11

2 Upgrading to version 9.1 Upgrading to version 9.1 Upgrading from version 4 or later to version 9.1 To upgrade Pexip Infinity software from v4 or later to v9.1: 1. Before starting the upgrade, we recommend that you use your hypervisor's snapshot functionality to take a full snapshot of the Management Node. If upgrading from v8, due to incompatibilities resolved in v8.1, ensure that you take a non-quiescing snapshot. This snapshot will be required should you later need to downgrade, or if the upgrade fails. You may also want to take snapshots of each Conferencing Node, although depending on the size and complexity of your deployment it may be easier to simply redeploy these from the Management Node in the unlikely event that this is required. 2. Reboot all Conferencing Nodes. There are two ways of doing this: o o Log in to the virtual hypervisor and reboot the machine using its processes. Connect to the Conferencing Node via SSH, login as admin, and issue the command sudo reboot. 3. Download the Pexip Infinity upgrade package for v9.1 from 4. From the Pexip Infinity Administrator interface, go to Utilities > Upgrade. 5. Select Choose File and browse to the location of the upgrade package. 6. Select Upgrade. There will be a short delay while the upgrade package is uploaded. After the upgrade package has been uploaded, you are presented with a confirmation page showing details of the existing software version and the upgrade version. 7. To proceed, select Upgrade. You are taken to the Upgrade status page, showing the current upgrade status of the Management Node and all Conferencing Nodes. This page automatically refreshes every 5 seconds. 8. When the upgrade completes, all systems will show a status of No upgrade in progress and have the new Installed version. If a Conferencing Node fails to upgrade, for example if it remains on a Waiting for calls to clear status, it should be rebooted. The upgrade process will then continue as expected. If you are using VMware snapshots for backup purposes, we recommend that you delete those snapshots after approximately two weeks, providing your upgraded system is operating as expected. This is because Virtual Machines, in general, should not run with snapshots over time. Upgrading from version 3 to version 9.1 To upgrade Pexip Infinity version 3 software to the latest version, you must first upgrade to version 6. To do this: 1. Download the Pexip Infinity v6 upgrade file (contact your authorized Pexip representative for the link to this file). 2. Follow the steps outlined above, but when asked to Choose File browse to the location of the v6 upgrade file. 3. Verify that the upgrade has completed successfully. 4. Download the Pexip Infinity v9.1 upgrade file. 5. Follow the steps outlined above, and when asked to Choose File browse to the location of the v9.1 upgrade file. For more information For full details on upgrading Pexip Infinity, see the Pexip Infinity Administrator Guide Pexip AS Version 9.1.a May 2015 Page 2 of 11

3 New features and improvements in this release New features and improvements in this release Version 9.1 This is a maintenance release. Version 9 Feature Infinity Connect Web App and desktop client enhancements Description Instant Messaging and chat Conference participants using a variety of clients can chat and share links with other participants. Supported clients include Microsoft Lync, Skype, and Skype for Business. Chat is also natively supported in Pexip's own Infinity Connect suite (Web App and desktop client). Microsoft Lync and Infinity Connect clients can also chat when calling each other directly via the Pexip Distributed Gateway. Content sharing All users can now share images and PDF content from any browser. Notifications and selective unlocking for a locked conference When a new participant attempts to connect to a locked conference using a guest PIN (or without a PIN if it is an open conference), the Hosts are notified that the participant is waiting to join. Hosts can then selectively either allow the participant to join the conference, or reject the request and disconnect the participant. While waiting, the participant is held in the conference lobby, where they will see and hear the waiting for conference host message (these items are customizable via themes). Conference control Host participants now have the options to mute all Guest participants simultaneously, and to disconnect all participants simultaneously. Role selection when adding participants You can select the role (Host or Guest) when adding a participant to a conference. New and improved design The Infinity Connect Web App and desktop clients have been redesigned with a new look and workflow to support the chat, content sharing and conference control features. This allows, for example, conference participants who are using a video conferencing endpoint, or a Microsoft Lync client, to use Infinity Connect to connect to the conference without audio and video, but have access to all of the advanced features. Simplified rebranding It is now easier to rebrand the styles and images used in the Infinity Connect clients. You can add a logo, a background image and set the colors to match your own corporate branding. Existing customers who have rebranded the Infinity Connect Web App will have to reapply their branding in version 9. Any textual changes to labels and messages will also have to be reapplied. As any existing branding on the Conferencing Nodes will be removed when upgrading, customers may want to prepare their updated branding prior to upgrading. Full details of how to rebrand the Web App are in Infinity Connect Web App Customization Guide. Outbound video streaming to Content Delivery Networks and public services Pexip Infinity can output a dedicated multimedia stream to enterprise CDN (Content Delivery Network) streaming and recording services such as Wowza, Adobe, VBrick, and Microsoft Azure, and to public streaming services such as YouTube. Any Pexip conference can be streamed as a live event to an unlimited number of viewers, and can automatically be recorded and stored for later consumption Pexip AS Version 9.1.a May 2015 Page 3 of 11

4 New features and improvements in this release Feature Device registration External policy Description Endpoint devices can register to Pexip Infinity Conferencing Nodes. This allows Pexip Infinity to route outbound calls to those registered devices without having to go via a SIP proxy or rely on DNS. Pexip Infinity's external policy feature allows call policy decisions to be taken by an external system, based on the data sources that are available to that system. This allows Pexip Infinity administrators to implement routing decisions based on their specific requirements. The external system can, for example, return VMR configuration data and jpeg avatars to represent participants in a conference, and decide the location for the placement of media on a per call basis. For more information, see Pexip Infinity External Policy Deployment Guide. Backup and restore Location-based call control enhancements Administrators can backup and restore configuration data via the Pexip Infinity Administrator interface. Pexip Distributed Gateway You can optionally configure Gateway Routing Rules to place the outgoing call from a Conferencing Node in a specific location (rather than from the node that received the call). This means that internal LAN-facing nodes and NATted DMZ-facing nodes can act as a gateway between internal and external traffic. Automatically dialed participants You can optionally configure an automatically dialed participant to be called from a Conferencing Node in a specific location (rather than from the node being used by the participant who initiated the conference). Microsoft Lync integration enhancements Per-location Lync MSSIP domain You can specify the Lync MSSIP domain per location. If configured, it is used instead of the global Lync MSSIP domain in outbound calls to Lync from Conferencing Nodes in that location. This makes it easier for service providers, for example, to support multiple Lync domains. Direct gateway calls into the AV-MCU Calls received via the Pexip Distributed Gateway can now be configured to call directly into an AV-MCU multiparty conference (both ad-hoc and scheduled calls hosted on AV-MCU). Audio-only calls When making audio-only calls from Lync to a Pexip Virtual Meeting Room, the participants in the VMR will only receive your audio. However, the Lync user will receive video and audio from the VMR. The Lync call can be escalated to video in the normal way by starting video, if required. Resizing shared application content When a Lync user shares an application (rather than their screen), additional black borders are now removed and the shared content is resized to cover the full dual video stream for receiving endpoints/clients. STUN servers Conferencing Node STUN servers You can configure the STUN servers used by Conferencing Nodes on a per location and per gateway rule basis. This enables a Conferencing Node to discover its public NAT address in deployment scenarios where the TURN server is not located outside of the enterprise firewall. Infinity Connect WebRTC client STUN servers You can configure the STUN servers to be used by Infinity Connect WebRTC clients when they connect to a Conferencing Node in a specific system location. In previous releases, Infinity Connect WebRTC clients always used stun.l.google.com. This STUN server address is included by default in v9. Customers upgrading from previous releases will have the stun.l.google.com server address added automatically to all existing system locations, resulting in no change to previous behavior. Join/leave notifications Themes now include the option to play audio tones to notify when participants join or leave a conference Pexip AS Version 9.1.a May 2015 Page 4 of 11

5 New features and improvements in this release Feature Additional audio notifications Description New audio files have been added so that: when the last Host leaves a conference, the remaining Guest participants are played a notification that their call will be disconnected shortly when there is just one participant in a conference, they are notified that they are the only participant. Preconfigured themes Layouts Pexip Infinity now ships with 4 preconfigured themes that are either with or without entry and exit tones, and that refer to the "#" key as either "the hash key" or "the pound key". The following enhancements have been made to layouts: Different layouts can be applied to a Virtual Meeting Room as well as a Virtual Auditorium. There is now an additional layout option that allows you to show the main speaker only, in full screen (with no thumbnails). For Virtual Auditoriums, you can elect to pin the presenter to the main video when they are sharing a presentation, regardless of who is speaking. Disconnect all participants from a conference Conference participant status Syslog server connection protocol Ability to disable SIP UDP Pexip Infinity Administrator interface improvements You can disconnect all of the participants from a conference via the Status > Conferences page on the Administrator interface. Conference participant status information indicates if the participant is currently on hold or if the participant is a streaming or recording device. You can choose the IP transport protocol to use UDP, TCP or TLS when configuring a remote syslog server. You can enable or disable incoming calls on SIP UDP separately from those over SIP TCP/TLS. Disabling SIP UDP can help reduce SIP spam. On new installations of version 9, SIP UDP is disabled by default. Upgrades from existing versions will have SIP UDP enabled (assuming SIP was enabled previously), and can be disabled in Global Settings. Local timezone The Pexip Infinity Administrator interface now displays times using the administrator's local timezone. (Log timestamps continue to use UTC.) New Call control and Users navigation menu options There are now separate menu options for Call control configuration (H.323 gatekeepers, SIP credentials, SIP proxies, Lync servers, TURN servers, STUN servers and Policy servers) and for User configuration (User authentication, Account roles and LDAP roles). Alarm enhancements You will now receive a warning when: your license is due to expire within the next 60 days the syslog server is inaccessible Also, you can view all alarm activity by searching the Administrator log for administrator.alarm. This will show a list of all Alarm raised and Alarm lowered events. Alarm raised and lowered events Automatic Conferencing Node deployment improvements You can view alarms that have since been resolved by searching the Administrator log for administrator.alarm. This will show a list of all Alarm raised and Alarm lowered events. Verify VM manager certificate When automatically deploying a new Conferencing Node on a VMware host, you can control whether the certificate presented by the VM manager is verified before the connection is allowed. Faster resource path selection There is an extra step in the automatic deployment wizard to select a datacenter. The resource path selection is then filtered by the selected datacenter Pexip AS Version 9.1.a May 2015 Page 5 of 11

6 Issues fixed in v9.1 Feature Support for G.719 Endpoint IP addresses removed from analytics events Description Support has been added for the G.719 audio codec. Endpoint IP addresses are no longer included in the information sent when usage reporting has been enabled. Issues fixed in v9.1 Pexip 4431/ 4362 In rare cases, a VMR with zero participants was displayed as active on the Management Node. This issue has been resolved in v The install wizard now has improved validation to reject invalid gateway addresses when deploying the Management Node If licenses are deleted and added in rapid succession, the licensing engine in the Management Node no longer fails to activate the newly added licenses Conferencing Nodes now gracefully handle API connections where the connection is lost after a client-side API call has been made, but before the response can be generated If a participant joins a PIN-protected, locked conference and enters the incorrect PIN during the first attempt, they are no longer redirected back to the PIN entry screen on subsequent PIN validation The media_stream fields are now included on the /api/admin/schema/ page Entry and exit tones (when included in the conference theme) are now always heard by all conference participants The +n indicator on a 1+0 layout now shows the correct number of other participants in the conference A call will no longer be placed to an Automatically Dialed Participant (ADP), if an existing manually-dialed call is already being placed to the same alias Administrators that sign in to the Infinity web interface via LDAP authentication, with roles that contain all permissions, can now view individual participants in conference history, and media statistics history for a given participant The Management Node administration web interface features now work if Internet Explorer is placed in Compatibility View. Microsoft 4563 Lync clients that are at the conference PIN entry screen are no longer delivered chat messages from the conference Lync clients in a call to a standards-based endpoint via the Pexip Distributed Gateway, may have been unable to share content with that endpoint if the Lync client had previously attempted to send an instant message to that endpoint. This issue is resolved in v Rarely, connectivity to remote Lync clients may have failed if the Lync client selects the remote ICE candidates on the Pexip Conferencing Node itself. Permissions for all ICE candidates are now set correctly If a Lync client joins a VMR as an audio only participant, starts video after joining, and subsequently stops video, the participant would disconnect after approximately 90 seconds. The disconnect is caused by Lync not properly responding to a re-invite sent by Pexip when the audio channel is lost. This issue is resolved in v Pexip AS Version 9.1.a May 2015 Page 6 of 11

7 Issues fixed in v VMRs or Gateways that are added to Skype clients now show up immediately as valid contacts RDP connectivity will no longer fail due to the Lync client mis-reporting the desired port for connectivity checks If a Distributed Gateway participant connects into a Lync AV-MCU, it will no longer receive the conference avatar back as main video participant if no other AV-MCU participant is actively speaking Stealing the presentation channel from an active Lync RDP session can no longer cause a failed RDP channel setup to that Lync client. Issues fixed in v9 Pexip 3924 Distributed gateway calls that limit bandwidth to and/or from a Lync client will now throttle Lync clients properly Gateway calls made outbound to Lync that fork back to Pexip nodes, thus creating an INVITE loop, will now be terminated to prevent full resource consumption Administrative logs of connected RTMP clients (web-based connections from Microsoft Internet Explorer or Apple Safari) are now reported as RTMP participants Audio support for AAC audio protocols has been improved for RTMP/Flash players IVR prompts for participants connected over RTMP/Flash clients will now play full audio clips; the issue that caused the beginning of the prompt to not be heard has been resolved Media participants included within the participant history now properly reflect any content channel The boot process of IPv6 configured Conferencing Nodes has been improved The reliability of a Management Node IP address change via the startup wizard has been improved in this release. Previously, attempts to change the IP address may have failed, forcing a re-deployment Media resource allocation has been improved. This resolves an issue that would previously cause calls to fail in very rare circumstances WebRTC connected participants will no longer experience downspeeding as aggressively as they did in the past; Infinity now takes a more judicious approach to reducing bandwidth in order to balance quality and speed of connection Audio codec selection on re-invite will properly take the connected bandwidth into account at the time of the re-invite rather than using the bandwidth originally signalled The v2 client-side API will no longer create an audio avatar icon at token grab; the icon will only be created upon call initiation In very rare circumstances, normal audio may have been detected as DTMF tones when connected to the Virtual Reception (lobby), causing invalid PIN entry errors to be displayed. Also, participants in a Pexip Distributed Gateway call may have occasionally heard a DTMF tone. These issues have been resolved in v Management Node deployment processes have been improved to prevent initial startup issues due to some services failing to start up Lip sync processes have been improved in this release The first participant to connect in a different location may have experienced a delay of up to five seconds in receiving video for the first time. Connection times are faster in v9 and endpoints should no longer show a black screen or "no incoming video" Pexip AS Version 9.1.a May 2015 Page 7 of 11

8 Issues fixed in v9 Cisco 2989 Incoming BFCP presentations received from Cisco MXP endpoints over an encrypted SIP connection occasionally failed to be decoded. This issue has been resolved in v API initiated conference transfers of Cisco E20s connected to encrypted calls on Pexip Infinity have been improved; media will now properly establish post transfer in all cases. Microsoft 4428 Skype clients running version 7.3 or later are more reliably connect to VMRs and Distributed Gateways in v If a Lync 2010 client is transmitting dual stream presentation, another client connected to the same VMR can now overtake the presentation channel When connected to AV-MCU, the Pexip Distributed gateway will now support the ability for either side of a gateway connection to assume ownership of the presentation channel Selfview of some web cameras as they appear in Internet Explorer 11 has been improved to prevent extraneous video fragments from appearing unnecessarily. This issue was present in selfview only and were not in the transmit stream If a presenter starts, stops, and re-starts a presentation at an extremely rapid rate, the newly transmitted presentation is now always displayed on the Lync client Locking and unlocking spotlight of endpoints connected into the Microsoft AV-MCU intermittently resulted in no incoming video from participants connected via the Pexip Distributed Gateway. This issue has been resolved in v Microsoft Lync clients can no longer share content to VMRs when participants remain connected into the Virtual Reception (lobby) and not actually fully connected into the VMR ICE controller behaviour has been improved in v9; this will prevent a re-invite failure that would rarely happen if video is lost to a Lync 2013 client due to packet loss If a Lync client stops and re-starts video when connected to a Pexip VMR, the restarted video is now properly decoded and displayed into the conference. RADVISION 3745 Quality improvements have been made to the RADVISION endpoints Calls to a RADVISION Gatekeeper made purely to an IP address no longer fail Pexip AS Version 9.1.a May 2015 Page 8 of 11

9 Known limitations Known limitations Pexip 4621 When automatically deploying a new Conferencing Node, the progress bar on the Deploying Conferencing Node page may stop at 96% even though the status indicates "Deployment succeeded". This is purely a cosmetic issue. Providing the status changes to Deployment succeeded it has succeeded, even if the bar only says 96% When dialing out to an RTMP URL, if the target of that RTMP URL is a CNAME record in lower case, and the target A records are in upper case (or vice versa), the lookup will fail Unpredictable system behavior may result from restoring a backup taken via the Pexip Infinity Administrator interface if, since the backup file was produced, any Conferencing Nodes have been added or removed, or the Management Node has been redeployed. We recommend that customers use their hypervisor's tools to create and restore full snapshots of the Management Node VM instead Conference Status may incorrectly report the instant messaging count as 0, even when messages are exchanged in a conference If a Lync client is receiving w448p main video resolution and participant names are enabled, the participant name for one of the participants connected into the conference will display unnecessarily small If a VMR with a duplicate Conference Alias is added using the Management API, the newly added VMR will take the aliases from the previously created VMR, rather than rejecting the addition as expected The display name used for dial out participants is not properly updated Changing the name of a system location can result in a loss of service and misleading usage statistics, as the name is used as a reference ID across the platform. If a location is renamed, all Conferencing Nodes should be rebooted The roster displayed on both the Management Node and the Pexip Infinity Connect client will incorrectly display two participants for any Lync client connected when content is being shared. This is simply a reporting error and does not affect system operation, capacity, or licensing If during a single conference a new presentation stream is started but fails to connect, the last frame from the previous presenter may be visible to the participants of the conference In very rare circumstances, a presentation that has been sent for a long period of time from a standards-based video endpoint may, at times, stall and thus prevent forwarding from one slide to another. The only way to resolve this issue is for the presenter to disconnect and re-connect to the conference. Please contact support@pexip.com if you experience this issue. Cisco 4565 In a VMR conference, a presentation to a DX80 may fail after a few seconds If an outbound call is initiated from Pexip to a VCS Expressway that does not inherently trust the remote Pexip node, and where the SIP TLS FQDN is configured on the Pexip Conferencing Node, the protocol is SIP TLS, and the endpoint registered to the VCS Expressway starts sharing content before someone else shares first, content will not be properly received by the Pexip Conferencing Node. This can be resolved by configuring an outbound neighbor relationship from the VCS Expressway toward the Pexip Conferencing Node, forcing the call to the VCS Expressway over the H.323 protocol, or having another participant share content first If the presentation channel already active from an MXP is taken by another connected participant, the MXP may not properly receive presentation content Pexip AS Version 9.1.a May 2015 Page 9 of 11

10 Known limitations Microsoft 4465 Audio may fail to resume if a Skype client behind a NAT mutes its microphone for a prolonged period of time while connected to a Pexip VMR Rarely, outbound calls made from Infinity to Lync clients registered to a Lync Edge server will fail. Re-attempting that same call should succeed without issue Outbound video calls made from the Lync 2010 client for ios may only connect using audio rather than automatically escalating to video as expected If a Lync 2013 for Windows client joins a VMR as a video participant, pauses video, ends video, then resumes video, the main camera stream will not properly resume. This issue can be resolved by stopping and starting video again Internet Explorer Web App clients may inadvertently not receive some messages sent over a chat interface. This issue can be resolved by joining via any other browser Plain text messages sent to Lync 2013 ios and Android clients receive a 415 error code in response If a Lync participant in an AV-MCU multi-party conference is presenting while another device joins the AV-MCU conference via the Pexip Distributed Gateway, the Lync presenter will see a "Someone has joined and can't see what's being presented or shared" notification. However, the gateway participant will be able to see the presentation. The notification will disappear after approximately 15 seconds Intermittent Distributed Gateway calls to the Lync AV-MCU will experience loss of video from AV-MCU when the remote client resumes a call previously placed on hold Participants connected via the Pexip Distributed Gateway into the AV-MCU may not receive presentation content from Lync participants. This occurs if the Lync user has a screen resolution where the width is an odd number of pixels, such as a resolution of 1437x Lync avatar and presence indications are not properly being presented on Lync for ipad clients Skype clients do not send DTMF tones to federated environments, therefore they cannot connect to PIN-protected Pexip Infinity conferences. The recommended workaround for conferences with Skype users is to use a semi-secure conference no PINs for guests, and to use the conference locking feature with admittance control via the Infinity Connect Web App If a standards-based endpoint transmits a dual stream presentation at a very low resolution, the transcoded presentation will be sent in native resolution to any connected Lync clients. This may create a sub-optimal experience depending on the PC screen resolution of the Lync end-user PC If a RADVISION H.323 endpoint joins an AV-MCU conference through the Pexip Distributed Gateway, video may be lost after a prolonged period of time Locking and unlocking spotlight of endpoints connected into the Microsoft AV-MCU via the Pexip Distributed Gateway may intermittently result in participants seeing the audio indicator instead of video Participant status of an AV-MCU may improperly reflect an old call ID within the Administrator and Support logs on the Pexip Management Node. This is a reporting error only and will not impact system operation If a Cisco VCS is used as call control between a Conferencing Node and a Lync 2013 FEP, only audio calls are possible. Lync FEP and Conferencing Nodes should be neighbored directly then audio and video calls will work as expected. The Lync 2013 client always uses VSR packets to negotiate the video mode even for RTVideo calls. However, Pexip Infinity assumes that Lync clients will always use video preference RTCP extensions to negotiate the RTVideo resolution and only looks for X-H264UC when processing VSRs Pexip AS Version 9.1.a May 2015 Page 10 of 11

11 Known limitations Polycom 4620 If you are using older Polycom CMA or RealPresence Desktop clients, G.719 audio may cause audio issues with low bandwidths (512k and lower). Please contact if you experience this issue Starting and stopping a presentation many times may cause a connected Polycom HDX endpoint from receiving the presentation stream. VMWare --- Disk access on ESXi 4.1 is typically considerably slower than on ESXi 5 and above. Tasks that require significant disk access, such as importing large conference configurations or writing many frequent log entries, may take over 5 times as long on ESXi 4.1 as compared to ESXi Pexip AS Version 9.1.a May 2015 Page 11 of 11