Ebbe and Oz talk Second Life in the cloud

Credit: Linden Lab / Amazon Web Services Inc.

It’s a huge effort. Right now the Second Life grid is a proprietary set-up in a hosting facility that we have customised for the purposes of what we’re doing – which made sense a decade and half ago. But these days, with the services from Google and Amazon and Microsoft with cloud infrastructures, it makes a lot of sense to shift our technologies to be on top of those cloud infrastructures instead of having our own.

Ebbe Altberg, VWPBE, March 15th, 2018.

That the Lab is working on moving Second Life to the cloud is becoming more and more widely known. First mentioned by Landon Linden (aka, Landon McDowell, the Lab’s Chief Product Officer) during his SL14B Meet the Lindens session, it was “officially” announced in August 2017 via a blog post.

Landon Linden, June 2017, talking about the Lab’s hope to move Second Life services to the cloud.

It’s a long-term project, which will extend well into 2019 (at least), building on a relationship with Amazon dating back to 2008, and which today both Blocksworld and Sansar (see: “Project Sansar”: an Amazon ECS case study), from which the Lab hope to gain a range of benefits, including – in time – perhaps the opportunity to offer a broader range of products at more comfortable (for users) price points.

There are some significant technology challenges the Lab faces with the move. However, progress is being made. Some non-user visible services are already running in the cloud, and more recently, the Lab has started preliminary testing with cloud-based simulators – although they are fair from ready for users to access, as Oz Linden outlined at the March 16th TPV Developer meeting:

We have actually run experimental regions on cloud servers, and it worked. There were some functional limitations that we have to do a lot of work to solve before we could begin to do regions that ordinary users can get to … It’s something we’re pursuing as aggressively as we can [but] I’m not even sure we have a sufficiently comprehensive view of the problems … some of them will only become apparent as we actually put things into production.

Oz Linden, TPVD meeting, March 16th, 2018 – full audio below.

Oz Linden, March 16th, 2018, talking about progress to date, and how things are likely to progress.

For the Lab, the benefits of the move to the cloud include things like a reduction in their capital expenditure  – no need to maintain their own dedicated hardware (or continuously update / replace it) within a dedicated operating environment. It also means they can more dynamically scale consumption according to needs – this could be beneficial for a number of the back-end systems within Second Life.

It turns it into less capital expenditure to have to buy all the equipment and doing all the maintenance on that. You kind-of pay for what you use; with Second Life [right now], once we’ve bought a piece of hardware, we have to sit on it whether it’s being utilised or not, whereas you can kind-of dynamically scale your consumption as necessary when you use something like AWS … which we believe will reduce costs for use and then ultimately, we hope to pass that on to customers.

Ebbe Altberg, VWBPE, March 15th 2018.

Once the transition has been completed and the Lab has had time to evaluate things, the move might allow them to offer a more varied land product – something again touched upon in Ebbe Altberg’s 2018 VWBPE address, and allow them to more extensively “re-balance” the revenue model – something that is also an ongoing project at the Lab.

We’re really thinking hard about the economic model of Second Life. We share a belief inside the Lab that land is quite expensive. so we’re constantly looking at ways to lower land prices and find other ways to find revenues. So I think you will see us try to shift from what I would say [are] high real estate taxes to more consumption taxes or fees to create an environment where it’s easy for people to create and own experiences, and we [the Lab] participate more in all the transactions that take place.

Ebbe Altberg, VWBPE, March 15th, 2018.

Given that land tier provides the lion’s share of the Lab’s revenue, this re-balancing is far from easy to say nothing of the potential for user outcry at any fee increases). Ergo, having better means to lower fees such as through reduced operating costs and a broader spread of more “affordable” products could – depending on the time frames involved – go a long way towards helping the Lab achieving that re-balancing goal.

So what might the move to the cloud mean for users? That’s hard to quantify at the moment, simply because the project has so far to go.However, some hints at what might happen have been offered.

For one thing – and on the subject of different land products – it might allow the Lab to offer two broad categories of region / server type; I’ll call them “always on” and “on demand”.

  • “Always on” would be simulators running 24/7 as with SL at present. These would be ideal for handling Mainland, large open spaces like Blake Sea and the larger, contiguous private estates. Such regions might have a similar type of fixed-fee tier cost associated with them as we have today (although not necessarily the same price points).
  • “On demand” would be simulators that are only active (and charged for) when in active use. When devoid of avatars, they are saved to disk and spun down. These types of region could be ideal for special events, or for private business / residential regions which don’t have any surrounding regions, and would only be charged for when avatars are present; once the last avatar leaves, following an appropriate pause, the region is saved, and the instance spun-down.

Such an approach has been alluded to by Ebbe Altberg:

Some experiences might want to have continuous persistence over time, and maybe that’s one type of pricing model, for an “always on” type of scenario. Maybe other will be fine with, “hey, I’m only using this for a few hours in a class a few times a week” or something. and if that can spin-up in a few seconds, and then I just need to basically pay for the time that I’m utilising it. Those could be potential options for us to explore.

Ebbe Altberg, VWBPE, March 15th, 2018.

Land offerings could be broadened in other ways. Again, as Ebbe Altberg indicated at VWBPE 2018, there might be high-performance, high-capacity, “upper tier” servers available for those needing them for specific uses (e.g. events need high concurrency levels or similar), sitting alongside more moderate, lower-cost servers for things like residential use.

More intriguingly, cloud hosting might even allow the Lab to more readily geo-locate simulators / regions with their physical world audience. Such regions wouldn’t necessarily have to be grouped together in-world, they are simply located a lot close to the user base that most frequently uses them, potentially improving performance for that audience.

Today we are located in the US, which means that people from Australia or Asia or Europe have to travel quite a ways, which is hundreds of extra milliseconds of latency. So if you want to have a very dedicated community in Australia or somewhere, we could maybe start to distribute our server infrastructure to be closer to where the actual customers of those regions are, which would make things more performant.

Ebbe Altberg, VWBPE, 15th, March 2018.

There will be more to come on SL and the cloud and the Lab provide further updates as the work progresses, and I’ll hopefully report on them as they are made public. In the meantime, and for those who haven’t waded all the way through the VWBPE 2018 video with Ebbe Altberg (and Brett Linden), or who don’t want to read either my transcript of that event or the bullet-point summary, here’s the audio of Ebbe’s comments on SL and the cloud:

Advertisements

2018 SL UG updates #11/2: TPVD meeting with Ebbe Altberg

Realm Of Light; Inara Pey, February 2018, on Flickr Realm Of Lightblog post

Update: the 43-bit viewer KDU issues has been updated based on feedback from Ansariel Hiller.

The following notes are taken from the TPV Developer meeting held on Friday, March 16th 2018. A video of the TPVD meeting is embedded below, my thanks as always to North for recording and providing it. Time stamps in the text below will open the video in a new tab at the relevant point of discussion.

This meeting was somewhat extended  – lasting 1 hour 30 minutes – as a result of the presence of Ebbe Altberg, Linden Lab’s CEO, who commented on some of the item of discussion that came up at his session at VWBPE 2018 (see my notes and transcript here), as well as more broadly discussing Second Life and Sansar.

SL Viewers

[0:17-2:20] There have been no further updates to the current release, RC and project viewers in week #11, leaving the pipeline as follows:

  • Current Release version 5.1.2.512803, dated February 23, promoted March 1 – formerly the Nalewka Maintenance RC – No change
  • Release channel cohorts :
  • Project viewers:
  • Linux Spur viewer, version 5.0.9.329906, dated November 17, 2017 and promoted to release status 29 November – offered pending a Linux version of the Alex Ivy viewer code.
  • Obsolete platform viewer, version 3.7.28.300847, May 8, 2015 – provided for users on Windows XP and OS X versions below 10.7.

General Notes

  • The Media Update RC viewer is unlikely to be promoted to release status in the immediate future, as it has some Windows 7 update issues which need to be resolved.
  • The Love Me Render viewer is making good progress, although it also has the Windows 7 problem.
    • [7:30-8:30] This viewer also has a KDU issue which can cause the 32-bit version of the viewer to crash when uploading textures larger than 512×512. One workaround for this until fix is obtained – depending on how long that takes – is for an older version of KDU to be used for 32-bit viewer versions.
  • Despite the issues with it (see my update here), the 360-snapshot project viewer is not getting a lot of attention.
  • Animesh project viewer is getting close to a possible RC release and the Animesh project close to a move to the main grid.
  • The Bakes on Mesh viewer has cleared LL’s QA, so a Bakes on Mesh project viewer for use on Aditi should be appearing soon.

New Viewer Caps

[3:51-7:04] The lab is introducing two new viewer caps they’d like TPVs to adopt quickly:

  • One will be used when the viewer first logs-in to read all of the deferred IMs received while the user was off-line, which are being moved from UDP delivery to HTTP in an attempt to overcome issues of off-line IMs failing to show.
  • The second is to read the correct set of abuse report categories from the server, so only valid categories are displayed within the viewer, allowing users to more correctly file ARs, rather than using invalid categories held viewer-side.

 

General Discussion with Ebbe

Highlights only – refer to the video for the full discussion.

Economic Model

[9:20-14:35]

  • The Lab is looking to try to pivot the SL economic model away from a heavy reliance on land fees, and then in time hopefully reduce the cost of land.
  • This will see a shift from land to revenue generation through fees in other areas.
  • The Lab cannot simply drop land fees and raise fees elsewhere, the two have to be balanced, so while the Lab is hoping to “aggressively” tackle pivoting revenue generation, they will also be cautious in making changes.
    • The hope of reducing land fees can be seen in the reduction in Mainland fees.
    • [10:50 via Oz and Linden] The Mainland price reduction has already seen a significant uptick in interest for abandoned Mainland, with support being “overwhelmed” with requests.
  • There is currently nothing planned for Private region fees, simply because the Lab has to be cautious around revenue.
    • It might be a case of (a) fee increase(s) elsewhere first, followed after a time by consideration of what can be done with Private region fees.
    • It is however, something the Lab would like to do.
    • The steps must be measured not only to safeguard LL’s revenue stream, but also so as not to upset the SL economy.
    • For this reason, the Lab will take a little time to measure the Mainland restructuring before they make other changes, so that they can more accurately measure cause and effect between different types of change.
  • These ideas were also discussed that the VWBPE session with Ebbe – see my transcript notes (with audio from that session) for more.

Sansar

[15:39-18:36] A general overview of Sansar – which is still is  Creator Beta – including the drive this year to gain an audience for Sansar, plus improvements to the VR aspect of the platform. Most of this is covered in my weekly Sansar updates. For Ebbe in particular, the Sansar team is at a point where he feels comfortable pivoting attention away from that platform and back to Second Life, including spending more time in-world.

Sansar and SL

[18:38-22:09]

  • A re-iteration that Sansar was never intended to be a replacement for Second Life.
  • Both products now have completely separate teams working on them
    • At VWBPE 2018, Ebbe indicated that the core SL team – engineering, development, operations, support – is “close to” 100 in number.
  • There is an area of overlap between the two products, but there are also very clear differentiators.
  • Proof that Sansar isn’t a replacement for Sl is the level and speed with which LL has continued to invest in SL (overhauling the viewer and simulator build mechanisms, bringing more performance and stability to the platform) and to add new capabilities (Bento, Animesh, Bakes on Mesh, EEP, rendering enhancements, etc.).

Moving SL to the Cloud

[22:09-32:00]

  • Progress is being made.
  • Experimental regions have been run in the cloud, and they worked.
    • There are a lot of functional limitations that must be addressed before regions users can access can be run in the cloud.
    • The regions did achieve a reasonably high concurrency level (precise number not given).
  • Much of what SL does natively – dynamically spinning-up a new set  of inventory management servers or a new set of log-in services, etc. – is similar in nature to a lot of what cloud service providers do, so a lot of the back-end work involved in moving to the cloud is taking what the Lab have, and adapting it to run within the infrastructure of the cloud.
  • It is a massive engineering undertaking that will take time.
  • Once completed, it is hoped operating SL in the cloud will allow LL to offer benefits to users, which might potentially include:
    • Reduced costs for regions that are spun-down and stored to disk when no-one is using them, should this be explored for Second Life
    • Ability for simulators and services to be more geographically based (e.g. simulators used largely by an audience in South America could be hosted in facilities in South America)
    • Ability to potentially have a broader cross-section of land product based on server types, with a broader range of performance / pricing.
  • It is hoped that, for the most part, users won’t be aware of services being switched from the Lab’s dedicated infrastructure to running within a cloud infrastructure
    • Some non-user facing services are already running in the cloud.
    • The work will be done progressively, and not a “flipping of the switch” for “everything”.
    • There is not end-date for the work. The Lab is approaching it as aggressively as possible, but there are a lot of technical hurdles to be cleared along the way, some of which will only become apparent as attempt are made to shift things and put them into production via the cloud.
    • To deal with potential issues / hurdles, it is possible that further ahead, there is a simulator RC channel “in the cloud”, while others are still running on the Lab’s own infrastructure.
  • Also see Ebbe’s comments from VWBPE 2018.

Upcoming New User Experience

[1:01:23-1:07:11] The Product Team is doing a lot of work with the new user flow, and are getting close to where they can start experimenting with ideas.

Part of this work involves a themed learning island reached via a new user clicking a themed ad which takes them to a themed landing page on the SL website, where they can sign-up, obtain an avatar in keeping with the theme, and are delivered to a learning island that also follows the theme.

This approach will be tested alongside the current on-boarding routes.

Interesting tidbits:

  • The lab spent over a year building a “fairly sophisticated” tracking system to gather data on new users and see how they are doing, i order to try to learn more about on-boarding / retention.
  • The Lab’s data / testing suggests new user retention is no better in welcome areas with greeters, than for those without greeters.
  • A test the Lab carried out using a (non-public) browser-based means of accessing Second Life from sign-up (no need to download the viewer) also did not achieve better retention than the “traditional” sign-up and download route.

Other Items

  • [32:10-34:07] Mirrors in SL: the inevitable discussion – and no, mirrors aren’t in Sansar!
  • [34:08-37:08] SL and VR: re-cap of why the VR viewer was dropped from Second Life – unable to maintain the comfort-level VR frame rates (90 fps). Also segues into a discussion of the Sansar / SL Edit mode differences (also see Ebbe’s comments from VWBPE 2018).
    • [37:10-37:40] Sansar benefits to SL: Oz confirms that some of the rendering work with the atmospheric shaders to improve SL’s appearance is leveraging lessons learned with Sansar.
  • [38:35-43:50] Texture caching: the project to improve the viewer’s texture caching is still very active, and once completed, the Lab plain to look at other aspects of how the viewer caches data.
  • [44:17-47:09] Linux Viewer: no real change from my last update.
    • TPV have the same problem as LL re: Linux developers.
    • LL would like to see more from the Linux community get involved.
    • Suggestion is for Linux users to try running the Windows viewer under Wine.
  • [47:10-49:25] OpenGL and GPU/CPU divide: discussion on updating SL’s OpenGL version, which is already under consideration at the Lab. Broadens into a discussion of modifying SL’s rendering capabilities to make more use of more GPU shader capabilities for calculation (rather than being reliant on the CPU), and the risks (to users) this entails (as many SL users don’t use more modern hardware with GPUs capable of taking the load).
  • [51:58-59:58] Community Gateways: Discussion on the Community Gateway programme and attracting users. Includes mention of partners, Place Pages, etc. Ideas raised seen as something that could be put to the SL Marketing team under Brett Linden.
  • [1:12:52-1:14:14] Feature Requests: when filing a Feature Request JIRA, it is better to keep the request focused on a single idea which can be easily digested. Multiple ideas should be submitted via separate JIRAs so that meaning don’t become confused / the JIRA becomes to complicated to understand, etc. Multiple JIRAs around related ideas can also be related via identifier.

Ebbe Altberg at VWBPE 2018: video, transcript and audio

VWBPE 2018 banner. Courtesy of VWBPE

On Thursday, March 15th, 2018, Linden Lab CEO met with a group of invited participant for the Above the Book events at VWBPE 2018. Lasting some 80 minutes, the conversational  Q&A session was livestreamed for its first hour to the main auditorium at the conference. Also in attendance at the event was  Brett Atwood (Brett Linden), Senior Manager for Marketing at Linden Lab, was also in attendance.

The full transcript of the session, including a number of “after stream” questions not in the official video is available here. When reading, please note:

  • The transcript notes don’t follow the chronological flow of the session, as I’ve attempted to group subjects by topic for more concise reference.
  • Audio of responses to questions are provided for each section of the transcript.
  • Questions are given in bold.
  • Comments from Ebbe Altberg are prefixed EA; comments from Brett Atwood are prefixed BA.

The video of the session is embedded at the end of this piece for reference and for those who would prefer to watch it. My thanks to the VWBPE stream team for providing it.

The Short Form

For those who prefer,the following is a bullet-point list of the core questions asked, gathered into the same topic headings as the transcript notes. Links are given to the relevant part of the full transcript for those who would prefer to read / listen to the complete answers / comments.

Note: due to the nature of WordPress internal text anchors and scrolling, you may have to scroll up a little after jumping to a specific topic, in order to see the question.

The Lab and Education

  • Education is, and remains, an important element of the lab’s thinking, both with Second Life and with Sansar. Jump to topic.
  • For educators interested in engaging with the Lab, the best way to do so is by reaching out to Brett Linden, who himself is an educator with 15 years experience (and who currently teaches a course as Washington State University) and who sees himself as a advocate for virtual worlds in education.
  • To help educators promote their endeavours, Brett suggests looking at SL Place Pages as a means to self-curate a web presence for in-world activities. Jump to topic.
  • The Lab is willing to work with educators on putting together documents / information addressing the value proposition, safety, security, etc., of using virtual space in education which can be used by educators to help inform their administrators and IT experts. Jump to topic.
  • The Lab is also willing to consider ways to make SL more attractive to the education market, preferably with a list of prioritised ideas from the education community, which can be compared with other requests for features / improvements, evaluated (technically and in terms of value / ROI, etc). Jump to topic.
  • It’s unlikely that LL will, in the near future, provide a dedicated resource to the educational / non-profile community. However, the recent re-alignment of the Second Life and Sansar team structures means there is now a dedicated SL marketing resource, lead by Brett, to help and support SL communities such as the educational / non-profit community. Jump to topic.
  • What has been the educational uptake with Sansar?

Product Accessibility and Ease-of-Use

  • Access for the disabled:
    • Sansar is too early in its development to offer much in the way of ease-of-access for the disabled.
    • The Lab would like to hear of specific instances where more could be done to help with accessing Second Life by those with disabilities.
    • Jump to topic.
  • What is the Lab’s position on speech-to-text, text-to-speech?
    • There are 3rd party tool already available, which the Lab might not look to directly integrate into their platforms.
    • However, major players like Microsoft, Google and apple are developing capabilities which the Lab might look towards integrating in the future.
    • Jump to topic.
  • Providing a plug-in architecture for SL / Sansar to allow third-party extensions (text-to-speech, etc.), has been suggested.
    • Not n the current roadmap, as SL client already provides a lot of open-source extensibility.
    • Possibly something TPVs could implement / contribute.
    • For some services, would have an associated cost involved, adding complexity, simply because third-parties charge for their services.
    • Jump to topic.
  • More broadly, Sansar has offered accessibility for people to be able to create and publish their own VR experience to the world without the need to develop their own platforms on which to run them. Jump to topic.
  • Will Sansar have easier access to content tools? Jump to topic.

Second Life: Development, Land Cost and Fees

  • The Second Life team numbers just under 100 people (engineering, support, operations, customer services, and not including finance, compliance, etc.). Jump to topic.
  • Key focuses for SL are: enhancing the platform, growing  the user base and trying to re-balance the economic model away from a reliance on land fees.
    • Costs of involvement in Second Life remain a concern for the Lab, which is taking a high-level look at trying to pivot revenues away from the land model potentially towards more of a transaction-based model, while at the same time trying to reduce at least some of the burden of land costs (e.g. the 2016 private region buy-down offer and more recently, the Mainland price restructuring).
    • Jump to topic
  • In time it is hoped that transitioning revenue away from land might lead to a further reduction in land costs to educators. Jump to topic.
  • Other areas of focus include making SL more scalable, secure and performant. Jump to topic.
  • The Lab will soon be publishing a roadmap announcement of improvements / enhancements to Second Life. Jump to topic.

Developing Second Life and Sansar

  • Linden Lab has more clearly split the Sansar and Second Life teams; there are no longer people “floating” between the two and working on both.
  • The Lab see a lot of overlap between the two products, with customers potential using both for different purposes.
  • However, each has its own potential markets / its own selling points for those markets, allowing them to co-exist.
  • The Lab believes that with 15 years of SL notwithstanding, the potential of virtual spaces has only been scratched, and both second Life and Sansar can do so much more.
  • Jump to Topic.

Second Life and the Cloud

  • Huge engineering effort, still in early stages, which will take “all of this year and then some” just to get transitioned and working.
  • Benefits to the Lab:
    • Reduced capital investment and maintenance costs.
    • Potential to dynamically scale consumption, rather than having simulator servers always on.
  • Potential benefits to users:
    • Potentially more flexibility in products and tiers, possibly defined by a mix of region use and use-case (e.g. high-powered, high capacity servers for popular events; more moderate, lower-cost servers for residential use).
    • Might add the ability to geographically locate simulators closer to dedicated audiences (e.g. regions serving South American communities could have their simulators running out of cloud locations in South America, reducing issues like latency for customers).
    • Not something that will happen immediately, as the Lab will need to transition services and then refine operations and investigate option.
  • Jump to topic.

Second Life and VR / Browsers

  • VR remains a challenge to implement in Second Life due to the high run-time frame rates required. As such, it is not something the Lab plans to pursue or re-deploy.
  • Streaming Second Life to a browser / mobile devices has been tried by third-parties (OnLive with SL Go and Bright Canopy from FRAME).
    • It’s not something the Lab is actively pursuing at present, so they are leaving it to third parties.
    • As the cost of cloud-base GPU technologies come down it might be something the Lab re-examines in the future.
  • Jump to topic.

“After Livestream” Questions

Note some !after stream” questions and comments are also addressed in the sections above.

  • Will the Lab bring back the Teen Grid as “the Education Grid”?
    • Not at this time, as it doesn’t fit with current plans to move away from dedicated infrastructure and to the cloud.
    • Might be something to be considered after the move to the cloud: presenting grids for unique audiences.
    • Jump to topic.
  • Can educators be given more control over their regions?
    • This would need indication from educators as to what is required (prioritised list again).
    • Jump to topic.

Sansar and Second Life in the cloud: LL speaks at AWS Re:Invent

Logos: copyright Linden Lab

It’s been a busy time at Amazon’s AWS Re:Invent conference, which closes in Las Vegas USA on December 1st. At the start of the event, Amazon announced the launch of their VR / AR development / publishing platform Sumerian (see: Sumerian: Amazon’s VR / AR app building platform for more).

Meanwhile, on November 28th, and potential of more interest to Second Life and Sansar users, the event saw Tara Hernandez, Senior Director of Systems and Build Engineering at Linden Lab, give a presentation covering Sansar and touching on plans for Second Life, entitled How Linden Lab Built a Virtual World on the AWS Cloud.

Most of the video delves into the intricacies of building a complex platform like Sansar and how Amazon’s products have empowered the Lab. As such, it does come across as quite a dry listen; however, within it there are some useful areas of focus which are worth noting.

For example, the early part of Tara’s presentation touches on some core truths about Second Life. Such as the fact it is a platform now 14+ years old, which started as an environment engineered almost down to the bare metal, taking advantage of what were, at the time, deep-seated optimisations in graphics and networking capabilities.

Over time, these have not only been layered upon almost organically over the years, but have also become – in Tara’s own words – “kinda ugly” in terms of trying to maintain and enhance. This monolithic, deeply rooted approach to the core elements of the platform is – along with the user-driven expectation than the user-generated content within the platform will not break as a result of changes to the platform – one of the major reasons  why “updating” Second Life isn’t simply a matter of JFDI, as might be thought.

Aspects such as compliance – another issue which is perhaps a lot more complicated than many might appreciate, given the complexities involved in running services like Second Life and Sansar, where the ability to cash out money adds a lot of additional regulatory overheads (visible and invisible from a user’s perspective) over platforms which only allow users to pay-in.

The video also reveals the depth of the relationship between Linden Lab and Amazon, which in the case of Second Life, stretches back to 2008, and which has encompassed the Lab’s other product, Blocksworld. In particular, it touches on Linden Lab using (and sometimes breaking!) Amazon’s more recent offerings, such as their ECS services, as a beta customer. This is something that Amazon has itself highlighted, featuring Linden Lab and Sansar in one of their own ECS use-case studies (see my article “Project Sansar”: an Amazon ECS case study, from January 2016).

ECS in fact drives almost all of the Sansar back-end, from the Atlas through to the store. In particular, the way in which the ECS application layer is used to present the Sansar Atlas, and manage the entire management of the experiences offered by the Atlas and their instancing, utilising Amazon containers (see 27:40-30:58).

How Sansar uses the Amazon ECA application layer to drive the Atlas & managing experience instancing (screen capture). Credit: Linden Lab / Amazon Web Services Inc.

What’s interesting here is not only the way in which Amazon’s services are being used, but in understanding what is going on from the moment a Sansar user clicks the Visit button in the Atlas, and the lessons the Lab are learning even now, as people use Sansar.

This latter point is itself of interest, as it helps to explain why Linden Lab opened Sansar up to wider audience in what seemed to many of us familiar with virtual space – myself included – to be a premature move. Simply put, they needed more of a flow of people moving through experiences to better judge how experiences can be more efficiently / effectively managed within the Amazon environment – spinning them up / down, instancing, optimising server use, etc.

In terms of Second Life, perhaps the most interesting part of the video can be found at 32:14-34:36, with a look at the recently announced attempts to move all of the Second Life service – including (eventually) the simulators, if possible – the cloud. Officially announced as a project in August 2017, but has been discussed at various in-world meetings such as the TPV Developer meetings.

Credit: Linden Lab / Amazon Web Services Inc.

In particular, the presentation touches on one of the major reasons for attempting the move: costs. Right now, Second Life is dependent upon hardware the Lab has to source and operate through a data centre. Updating this hardware, and the underpinning infrastructure  – network, fibre, rack space, etc., – requires continuous and high levels of expenditure (even allowing for re-purposing / write-down of old equipment).

There are also limits, as touched upon in the earlier part of the video, on what can be done within specific areas of Second Life support and maintenance. For example, Tara specifically mentions the core database services (which have been subject to numerous issues over the last year plus). While recovery times for these services has been halved – from three hours to 45 minutes – it is still a considerable outage period from the users’ perspective, and one difficult to bring down further.

Second Life in the cloud – challenges. Credit: Linden Lab / Amazon Web Services Inc.

Thus, an attempt to move Second Life to AWS could resolve a lot of issues for the Lab, and potentially allow them to leverage lessons learned with Sansar together with the capabilities of newer services – like ProxySQL – to further update and improve SL. It might also allow the Lab to move their database operations away from MySQL to more robust products, again following Sansar’s lead.

The shift of a platform from being data centre centric to cloud based is obviously non-trivial, and involves considerable challenges, some of which are outlined by Tara (above). However, from the comments she makes, she is anticipating possibly a dramatic level of progress over the next year. If so, it could be an interesting twelve months.

With thanks to Dassni – The Mesh Cloud for the Twitter pointer to the video.