Is the Hero Engine responsible of the bad performance of ESO?

killmove

Alcast wrote: »

zaria wrote: »

Alcast wrote: »

ESO uses a heavily modified version of the Hero Engine. Sure it is modified but the core issues are still present and is one of the main reasons this game has so many performance issues.

Unsure how much time ZOS invests into fixing issues or trying to improve things. If its heavily modified I guess they can call it their own now but then I do not understand why they would be happy with the current performance of their own engine.

It might use an heavy modified version, that is hard to prove without looking into the libraries and how the game executes.

However an huge part of performance problems is that so much calculations run on the server, you could have much larger fights in cyrodil on launch.
However insane amount of cheating and boting had them move most to servers overloading them.

Yes true, for me the game engine and network server/client side all go one in one. I am no dev and have no coding experience so maybe that might usually be seen as two things and not one.

Maybe ZOS should move most of the calculation to client side and add an anticheat (watcher) soft to monitor the client.
If cheat is detected then connection to server is cut.

Recremen

As with all such questions, we can't actually say anything definitive unless we know the actual implementation details. Lacking such, we can't really talk intelligently or truthfully about the engine itself, and instead have to settle for loose generalities (many players in one area leads to poor performance) and comparing to other products that use this engine. Even then, the engines themselves are getting heavily modified by each dev studio anyway, so it's still kind of a lost cause trying to pin blame on the engine. Unless maybe someone here can give me the big O for players firing off AoE skills in a Cyrodiil server and the exact engine subroutines involved, I don't think this is going to be a productive and factual conversation.

xaraan

jcm2606 wrote: »

Regarding whether ESO uses the Hero Engine, this is also partially an argument of semantics. If we assume that Zenimax started out using Hero Engine, then started heavily modifying it to better suit the game's needs, could the result be considered Hero Engine?

I say this because this is likely what happened. Zenimax used Hero Engine as an early prototyping tool, then slowly started chipping away at it, replacing existing code and systems with their own, until the engine became an amalgam of Hero and in-house.

Considering this is the first MMO under ZOS' belt, I sincerely doubt they dropped Hero Engine entirely and made a completely custom engine. MMO's are very massive and very complex projects, with lots of moving parts to screw up, so if a first-time MMO developer wrote their own engine in-house without outside help (nobody in the Zenimax Media chain has MMO experience outside of ZOS), it'd be a miracle to even have running, let alone survive for many years without a complete meltdown (ESO has had many bumps along the way, but it hasn't taken a complete nosedive in server stability, yet).

This is the best response to the whole discussion.

Chances are they didn't just create an engine from scratch on top of another existing engine they used to framework the game, but rebuilt much of the Hero Engine and keep the splash screen and more because they can't truly claim the engine is original since it might be legally argued that they copied concepts or code from the engine they were working from. So both sides can be right in their own way - zos built much of the engine from the ground up, but probably borrowed a lot of the same architecture and ideas from the Hero Engine as they did and cannot legally claim to have created an engine from scratch ideas.

That being said, I do think some of the issues come from the engine. A lot of companies keep tinkering with engines to make them do all sorts of amazing things they were not originally created to do, and sometimes you will hit a roadblock or run into limitations because of that.

idk

killmove wrote: »

because i found this video on youtube speaking about SWTOR (i don't know or play this game).

https://www.youtube.com/watch?v=kO-8ugqw_Kg&t=1191s

it seems the game has similar issue with performance as ESO

I only watches the first 5 minutes of long winded answers but it was the management of SWTOR that killed it. SWTOR has not experienced the server performance issues ESO has. I played it for a few years after launch and would go back for new raids until they stopped building raids.

jircris11

UPrime wrote: »

Complex combat that requires a lot of resources along with a lot of people in a tight spot. The reason is that the skill mechanics are too complex for that many people. Games that have a lot of people have much simpler skills. I don't mean the effects. But what the skill does on the server.

FINALLY some one who knows what they are talking about. But honestly lag has many causes from server,coding to your OWN net work,conecton,speed even net card. I get tired of ppl across the ocean complaining about lag, sadly it's due to distance. In na I see people complain about it but personally I'm at 80-100ms IN pvp.

Audrena

jircris11 wrote: »

I get tired of ppl across the ocean complaining about lag, sadly it's due to distance. In na I see people complain about it but personally I'm at 80-100ms IN pvp.

Actually, those of us "across the ocean" are getting epic latency and ability lag mainly thanks to ZOS's insistence on using Akamai to scrub each and every packet in case of DDOS.

Using that Akamai service results in connections from Australia being routed through Hong Kong and the Netherlands before finally arriving in Texas, adding insult (massively long round trip) to injury (the extra latency already introduced by the Akamai process itself).

Other companies running online games only use DDOS mitigation services like this when they are actually being hit with a DDOS.

ZOS uses it 100% of the time, presumably because they think of it as a "set and forget" solution, regardless of how badly it hurts the game's performance.

The engine seems fine to me. I'm running a nearly decade-old i5 CPU (with a modern GPU) and technical performance is fine. Network performance? That's not the engine, that's a service that was never designed for use in real-time online gaming getting in the way.

Xogath

There is very much something wrong with the optimization of the game and how it utilizes hardware.

There is absolutely no reason why a DX9 application should be performing so poorly given the hardware most people have today.

I have an 8086K and a 1080Ti, with 16GB of memory clocking in at 3200MHz.. I can run Witcher 3 on ultra all day long without any performance issues; and that's 2k resolution @ 120Hz no less. If I dropped it down to 1080p, I could push the boundaries even higher.

But ESO? The second I step in to a major city, my FPS tanks in to the 30 range. Sure, I can sustain 100FPS in a cave or my house, maybe even a dungeon with a handful of folks.. but I'm not sustaining the performance I should be sustaining at all.

As I've said before, I'd give up a year's worth of content for these issues to be fixed. The DX10, 11, or 12 API could be implemented for those of us who have the hardware that support it, and it would likely help.. some.. but the core of the problem lies in the optimization of the game, the engine, and probably messy, inefficient code at every avenue.

What most people don't understand is that from a development standpoint, where time is literally money, developers are forced to get things done as quickly as possible, so long as they're stable.. this often leads to very sloppy, lazy ways of doing things.. simply because those involved have no choice--they have strict deadlines to meet, because time means money.

That said, now that the game is actually off the ground and bringing in consistent profit, it's time for a systems overhaul with performance in mind.

Outside of that, the servers also need some work.. or hardware upgrades. The network lag in Cyrodiil is so bad that after acquiring Caltrops and Vigor on my Nightblade, I've never returned. My normal latency of 80ms or so instantly triples, or higher, as soon as I step in to Cyrodiil.. and that is no exaggeration.

There's no reason for that.. period. We pay for a stable playing environment, and we're being provided with something barely stable at every corner.

I'm not some crazy person trying to play the game on a 80" OLED screen hanging on my wall at 4K or higher resolution. I have a 2K, 144Hz monitor on my desk with hardware that is more than capable of playing games at that resolution and refresh rate (I cap it at 120Hz thanks to nVidia, though), because I do just that with other games all day long.

But I like ESO. I want ESO to do well.. it's a fun game; it's one of the best MMORPGs on the market currently. And it has the potential to stick around for a LONG time.. it honestly does.

But the exceptionally poor performance of the game, both in network and frames per second, is spreading like wildfire and something is going to have to be done about it.

Oh, and to the guy who talked about creating a custom profile for ESO to fix the performance problem.. both my CPU and GPU are at or above 75% utilization while ESO is running.

It is absolutely the game.

FlyingSwan

Xogath wrote: »

There is very much something wrong with the optimization of the game and how it utilizes hardware.

There is absolutely no reason why a DX9 application should be performing so poorly given the hardware most people have today.

I have an 8086K and a 1080Ti, with 16GB of memory clocking in at 3200MHz.. I can run Witcher 3 on ultra all day long without any performance issues; and that's 2k resolution @ 120Hz no less. If I dropped it down to 1080p, I could push the boundaries even higher.

Oh, and to the guy who talked about creating a custom profile for ESO to fix the performance problem.. both my CPU and GPU are at or above 75% utilization while ESO is running.

It is absolutely the game.

Interestingly, I have the same rig as you (but the 8770K instead of your 8086K and 32GB RAM) but my CPU runs at about 35% most of the time (across all cores) and my GPU about 40%. I also run 2k @ 144Hz G-Synced; there is no point having 4k on a desktop screen, you can't see the difference from 2k but it will of course consume more GPU bandwidth.

When busy my CPU usage may spike to say 70% and GPU to 50%, but even though I run Ultra graphics settings, the game refuses to push either component at all. Most modern games can certainly max out my 1080Ti, as the CPU is easily powerful enough to feed it graphics data at a high enough rate to keep it very busy, but ESO is just a badly optimised game and always has been.

Even if they move to Vulkan or DX12 the game will need significant recoding to work effectively with those more parallel APIs, but I don't see ZOS putting in effort to do that. They make their money in the Crown Store, so 'unseen' coding that cannot be monetised is something they don't care for.

Caligamy_ESO

Wait wait wait.. hold the press...

Are you guys saying ZOS has lied to us before?!?



vvvvvv

Cronopoly

killmove wrote: »

Just wandering why performance is so bad.

Alcast wrote: »

snip

This conversation is typically talked about in the highest level IE NAIVE above what is necessary in most threads to discuss a games performance. My job experience for over 21 years as Datacenter Operations / Architecture for a fortune 500 company.

We know so little of ZOS internal mechanics, and talking to the Engine can be irrelevant in this state. There are some factors that apply to ANY game engine middleware. I'll talk to a couple important items and how their setup and maintenance can affect performance regardless of the engine.

Your Online Character's SESSION on the "Super Server" for the Zone you are in should represent your characters digital state in the world. IE it will have serialization (Binary code) for all your character Stats, Clothing, Actions, In-combat, out-of-combat, Location, Interacting with World Object "X" etc... For you and everyone in the same Zone, all these PLAYER SESSIONS are side by side in the "Servers" memory.

[CPU aspect]
Now if no one was doing anything but standing there, the middleware software would have no interactions between player sessions per Clock Cycle. That said ANY game server can handle and complete a certain MAXIMUM number of Player session to player session Interactions cleanly per Clock Cycle. When there are too many Interactions than what the hardware/software can complete cleanly, they get queued up .... we see this as lag especially in PVP.

Of course, the games Engine Engineer(s) will try to code certain interactions to use as little CPU as possible, so that when they are spammed like in PVP the Actions CPU% stay's as little as possible time wise, allowing more transactions per cycle... Obviously If there are waaaay too many transactions, some queued up transactions "Should" get Timed-Out and completely disregarded by the game engine, yes we see this in PVP daily..."I pressed it and nothing happened" >_<

For any Game Engine Architect, this is his/her world. They live and breath keeping this transaction queue as clean as possible....In a perfect world if they are given the resources to do so...

[Database]
One of the things that can degrade it is the Database. Doesn't matter if Oracle, MS SQL etc. For any modern living Server, the Database, whether local to the physical machine or attached by Fiberoptics / 10Gig Networking being clean and up to date on maintenance is the end all be all. Over time a transactional type Database can degrade, and pathways to your data can get old. "Stats" need to be run during maintenance, Cache needs to be optimized for the new data etc. Many threads have been voiced about Game Server's performance for peak PVP encounters, just wanted to provide some info so some may have some clue as to what their focus should be on...

[Networking End to End]
I'll keep this one short. Most of you have no idea how many Network "Hops" your data goes through to get from your PC / Xbox/Playstation to ZOS datacenter. In many cases your network session will go through 20+ physical machines from your ISP, to other ISP's, Clouds, and several Firewalls along the way to it's end to end destination. 95% of which is outside ZOS Datacenter. Please keep that in mind when someone mindlessly says "the ZOS Network is a problem..."

Regards,

Cronopoly

AlienatedGoat

I think a significant reason for the poor performance of the game is much simpler.

Many of the developers that programmed the game starting from 2007 are no longer at ZOS.

Anyone who is or has been a programmer knows that it can sometimes be very difficult to work with someone else's code, especially when you don't have them accessible for reference.

Keeping that in mind, it's not hard to imagine the crazy amount of work that ZOS programmers have to do to track down bugs in code that could be up to 12 years old now.

zyk

I doubt any game includes branding from a different company in their splash screen unless they are legally required to. The presence of the HeroEngine logo (which I think is long gone?) was only evidence of a legal agreement.

It's completely conceivable that ZOS received a discount on their HeroEngine license by agreeing to include their logo, even if they did not use any code.

It's also conceivable that Idea Fabrik would would exaggerate the extent HeroEngine might have been used in ESO because it's a high profile game that promotes their engine. Even if HeroEngine's involvement in ESO is irrelevant, as COO of Image Fabrik, it is Sarah Grant's job to promote that as much as possible.

If HeroEngine exists in ESO in any meaningful way, it should be apparent to outside observers willing to delve into its libraries and processes.

I have no love for Matt Firor, but he was extremely clear:

so while we were prototyping the game on HeroEngine, we were simultaneously developing our own client, server, and messaging layer that were specifically designed with ESO in mind. Think of HeroEngine as a whiteboard for us – a great tool to get some ideas in the game and start looking at them while the production engine was in development.

What's precisely true? It doesn't matter. ZOS owns all of ESO issues, regardless of their possible origins. Scapegoating HeroEngine or Akamai doesn't accomplish anything.

Rudyard

Xogath wrote: »

As I've said before, I'd give up a year's worth of content for these issues to be fixed.

Yes, I'd rather this get fixed than see any other update in 2019!

killmove

I hope 2019 will be the year ZOS fix ESO many performance issues.

If this is made then ESO+ will be worth the money.

FlyingSwan

Cronopoly wrote: »

[Database]
One of the things that can degrade it is the Database. Doesn't matter if Oracle, MS SQL etc. For any modern living Server, the Database, whether local to the physical machine or attached by Fiberoptics / 10Gig Networking being clean and up to date on maintenance is the end all be all. Over time a transactional type Database can degrade, and pathways to your data can get old. "Stats" need to be run during maintenance, Cache needs to be optimized for the new data etc. Many threads have been voiced about Game Server's performance for peak PVP encounters, just wanted to provide some info so some may have some clue as to what their focus should be on...

Let's be clear though, the above type of database has been consigned to the history books and no one deploying an app or back-end infrastructure in the last few years will have deployed this type of database solution. It's all in-memory database (IMDB) now, even for the very largest apps and services that we all use everyday. This has mitigated the database being the pinch point that you rightly note, plus lowered TCO because we no longer live in a world where we pay DBAs and server engineers to manage old technology such as clusters ('pets'). It's all IMDB and continuous deployment in a software defined data center nowadays ('cattle'), even for the very largest services in the world, e.g. Google, Amazon and O365.

However, ESO is an app built with a legacy mindset - its development started in the old times of IT where DBAs could still get jobs and people still cared about hardware - and because a move to a modern architecture would also require costly business transformation (training, new operating model etc.), I can imagine ZOS will simply keep papering over the cracks of their legacy deployment, because a business case for an infrastructure refresh is not as sexy as one for shiny toys in a cash shop that can be shown to directly generate revenue. It's for these more costly to solve reasons, that even a much asked for move to Dx12 or Vulkan will not solve the endemic problems in a game that was clearly developed with what is now an outdated approach to technology.

And we can't really blame ZOS for this old hat approach to IT, the type of architectures we take for granted in 2019 were only just starting to overthrow the legacy model when ESO was released in 2014, so when they started to develop it, what we now consider legacy (servers, RDBMS, patching cycles etc.) would have been the status quo.

killmove

Looks like ZOS bet on the wrong horse from the start.

Grimm13

What about the Havok Physics and Script packages used in ESO?

Havok Script is discontinued, so has these old versions become buggy with latest releases of Lua?

Had looked up on the Havok site the listed games using their products. Wanted to know what Havok was since it's splashed up on load screens in most mmo's I play. What I found ESO uses the two but does not say the versions used.

Possible the Havok Physics is still the original used to design the game. Or has old code just been thrown in with the new engines without rewriting into a more compact updated code that the new releases use? Really not my wheelhouse so looking to understand how it might relate. Cursory look would suggest that the Physics would be a driving factor in combat as it handles collisions.

killmove

I don't understand why don't take the time and put the money to adress all those issues.

Look at FFXIV at launch it was a big mess but SE took the time to rework it with FFXIV ARR. Look at it now, it's a pretty huge mmo with almost any performance issues.

Grimm13

BTW, the following is taken from the Summerset credits. HeroEngine is still listed there. My question about the Havok also still stands, so from the credits it would suggest that it is a 2010 version?

(Edit: The Fork Particle, has it replaced the Havok? So old Havok physics code is been ported into the Fork Particle? Seems that both do the same thing. If this is the case it still seems clunky and better to have rewritten into newer code.)


Additional Software

Powered by Wwise © 2006-2018 Audiokinetic Inc. All rights reserved.

Awesomium © 2012 Khrona LLC. All rights reserved. Awesomium is a trademark of Khrona LLC.

Copyright © 2012. The Chromium Authors. All rights reserved. Copyright © 1995-2009 International Business Machines Corporation and others. All rights reserved. This software is based in part on the work of the Independent JPEG Group. This software is copyright © 1991-1998, Thomas G. Lane. All Rights Reserved except as specified. Copyright © 2000-2002 Glenn Randers-Pehrson.

Daniel Greenberg and Associates, LLC.

Uses Havok™: ©Copyright 1999-2010. Havok.com, Inc. (and its Licensors), all Rights Reserved. See www.havok.com for details.

Portions of this software utilize Speedtree® technology (©2009 Interactive Data Visualization, Inc.) Speedtree® is a registered trademark of Interactive Data visualization, Inc. All rights reserved.

Powered by HeroEngine, an Idea Fabrik technology, www.heroengine.com. Idea Fabrik and HeroEngine are registered trademarks of Idea Fabrik Plc. All rights reserved.

Uses Lingoona® Grammar Engine. © 2009-2018 Lingoona GmbH. All Rights Reserved.

Futura® Futura is a registered Trademark of Bauer Types.

Consolas™ Consolas is a trademark of The Monotype Corporation and may be registered in certain jurisdictions.

Univers™ Univers is a Univers is a trademark of Monotype GmbH registered in the U.S. Patent and Trademark Office and may be registered in certain other jurisdictions.

This software uses fonts produced by FONTWORKS Inc. FONTWORKS, and font names are trademarks or registered trademarks of Fontworks Inc.

Uses Simplygon™, Copyright © 2015 Donya Labs AB

Uses Fork Particle effects technology. © Copyright Fork Particle, Inc.

FaceFX software used for facial animation. © 2002-2006, OC3 Entertainment, Inc. and its licensors. All rights reserved.

Uses Bink Video. Copyright © 1997-2018 by RAD Game Tools, Inc.

Uses Granny Animation. Copyright © 1999-2018 by RAD Game Tools, Inc.

©2018 ZeniMax Media Inc. The Elder Scrolls Online: Summerset™ developed by ZeniMax Online Studios LLC, a ZeniMax Media company. ZeniMax, The Elder Scrolls, Summerset, ESO, ESO Plus, Bethesda, Bethesda Softworks and related logos are registered trademarks or trademarks of ZeniMax Media Inc. in the US and/or other countries. All Rights Reserved

Cronopoly

FlyingSwan wrote: »

Let's be clear though, the above type of database has been consigned to the history books and no one deploying an app or back-end infrastructure in the last few years will have deployed this type of database solution. It's all in-memory database (IMDB) now, even for the very largest apps and services that we all use everyday. This has mitigated the database being the pinch point that you rightly note, plus lowered TCO because we no longer live in a world where we pay DBAs and server engineers to manage old technology such as clusters ('pets'). It's all IMDB and continuous deployment in a software defined data center nowadays ('cattle'), even for the very largest services in the world, e.g. Google, Amazon and O365.

However, ESO is an app built with a legacy mindset - its development started in the old times of IT where DBAs could still get jobs and people still cared about hardware - and because a move to a modern architecture would also require costly business transformation (training, new operating model etc.), I can imagine ZOS will simply keep papering over the cracks of their legacy deployment, because a business case for an infrastructure refresh is not as sexy as one for shiny toys in a cash shop that can be shown to directly generate revenue. It's for these more costly to solve reasons, that even a much asked for move to Dx12 or Vulkan will not solve the endemic problems in a game that was clearly developed with what is now an outdated approach to technology.

And we can't really blame ZOS for this old hat approach to IT, the type of architectures we take for granted in 2019 were only just starting to overthrow the legacy model when ESO was released in 2014, so when they started to develop it, what we now consider legacy (servers, RDBMS, patching cycles etc.) would have been the status quo.

I hear you and fully agree. I do wonder where ESO is on a Sunrise / Sunset model for their hardware. If in a consigned Datacenter are they in a model where they can upgrade to newer hardware, as some new "shiny" is typically available every couple years or so if on Leased Hardware...(pure speculation).

Or did ZOS put their specific 4 year old Hardware in a Datacenter (theirs or a Colo) and are riding it 'til it dies...?

I've seen both models, and others to be sure. It just a tad unsettling with the current player caps in PVP today that the peaks cannot be aligned so that the massive lag we see does not occur...as frequently.

I also wonder if required cheat prevention (server side validation) basically just sucked the life out of any performance profiling to the point where we'd need massively faster server side architecture to handle the transactions, which puts ZOS in this pickle of peak performance lag...but 90% of PVE is just fine....

Regards,

Cronopoly.

P.S. Good to converse with a knowledgeable person every now and then lol.
You of course are correct on the in-memory model for anyone doing current apps today especially those that are not moving Billions of $$$ dollars a month in their apps... A current day game shard could crash completely and some in memory real-time data lost and no one in mgmt is going to freak out too bad. My example was for Apps where there had to be one or two hands on solid ground at all times, massive hardware redundancies, SRDF of critical data to IBM (BCRS), Mimix, Timefinder on Storage attached to the database etc...Getting carried away ofc...

FlyingSwan

Cronopoly wrote: »

FlyingSwan wrote: »

Let's be clear though, the above type of database has been consigned to the history books and no one deploying an app or back-end infrastructure in the last few years will have deployed this type of database solution. It's all in-memory database (IMDB) now, even for the very largest apps and services that we all use everyday. This has mitigated the database being the pinch point that you rightly note, plus lowered TCO because we no longer live in a world where we pay DBAs and server engineers to manage old technology such as clusters ('pets'). It's all IMDB and continuous deployment in a software defined data center nowadays ('cattle'), even for the very largest services in the world, e.g. Google, Amazon and O365.

However, ESO is an app built with a legacy mindset - its development started in the old times of IT where DBAs could still get jobs and people still cared about hardware - and because a move to a modern architecture would also require costly business transformation (training, new operating model etc.), I can imagine ZOS will simply keep papering over the cracks of their legacy deployment, because a business case for an infrastructure refresh is not as sexy as one for shiny toys in a cash shop that can be shown to directly generate revenue. It's for these more costly to solve reasons, that even a much asked for move to Dx12 or Vulkan will not solve the endemic problems in a game that was clearly developed with what is now an outdated approach to technology.

And we can't really blame ZOS for this old hat approach to IT, the type of architectures we take for granted in 2019 were only just starting to overthrow the legacy model when ESO was released in 2014, so when they started to develop it, what we now consider legacy (servers, RDBMS, patching cycles etc.) would have been the status quo.

I hear you and fully agree. I do wonder where ESO is on a Sunrise / Sunset model for their hardware. If in a consigned Datacenter are they in a model where they can upgrade to newer hardware, as some new "shiny" is typically available every couple years or so if on Leased Hardware...(pure speculation).

Or did ZOS put their specific 4 year old Hardware in a Datacenter (theirs or a Colo) and are riding it 'til it dies...?

Unsure if this is known, but I think that development started in 2007, so some key architectural decisions were probably made back then which tied them to their forward path. At that time, dual core CPUs and DX9 were the norm for example. So a decision might have been made about the threading which was reasonable at the time so we got stuck with this 'main-thread' thing which has been the bugbear of the client side until the recent multi-core improvements, which do actually seem to have helped me, it has to be said.

But clearly back in 2007 they would have been developing for big physical servers, n-tier architecture, and this whole monolithic 'mega-server' idea. Now they might go with a micro-services architecture so if the Crown Store fails they can take that out of play and redeploy, while we still play the game, same if they get an issue with say the inventory service etc., thus preventing all these maintenance periods affecting all elements of the game for protracted periods. Also would help with performance most likely, by decoupling services so they are not tightly bound and therefore not all threatened by one non-performant element. But you know, hindsight is 20-20!

I get pissed off with the flakiness of the game from time to time, but we're really living with something that's over 10 years old, not 4, so it's largely expected. As you say they may be on leased gear that does get updated, but even if so, architecturally they are constrained for sure, and a complete re-architecting of the game would introduce so much risk of outage and new bugs that it won't ever happen.

I am sure there was an article about how it all hung together that someone posted way back, but it may be my fertile imagination!

MaxwellC

Honestly it can't be a server issue and this is coming from a guy with a CCNA in data + working on cloud networking soon. My mate who works extensively with the unreal engine making games told me it's definitely their engine unless they're using a monitoring program that heavily impacts the game due to authentication checkings but that is only based on how often it is and what information is being pulled up and how it's stored/ if it's stored/ policies/etc lol.

killmove

Can we imagine ZOS doing a whole rework of the game by rebuilding and using a new engine (with all 2019 technologies) ?
they can make it so previous code (from actual engine) can be copied/pasted in the new one. that way they can keep all maps, models, etc. available in the game.

Don't know if we call this a remaster or a remake, I'm not a developper.

FlyingSwan

killmove wrote: »

Can we imagine ZOS doing a whole rework of the game by rebuilding and using a new engine (with all 2019 technologies) ?
they can make it so previous code (from actual engine) can be copied/pasted in the new one. that way they can keep all maps, models, etc. available in the game.

Don't know if we call this a remaster or a remake, I'm not a developper.

In development terms a new engine with new techs would be termed 're-architecting' or 'rebuilding', depending on the extent of the change. Both are significant application changes and cost money and introduce risk, but the outcomes can be worth it in terms of lower cost of ownership, higher availability, improved performance. It's the sort of thing businesses do all the time in a controlled fashion to retain a technology edge.

The only MMO where I know this was done was FFXIV (but I may be wrong), but the thing was dead on arrival as it was just such a mess. ESO, much as we get frustrated with it, seems to be well and truly alive, so any re-architecting risks alienating those people who are using it and getting enjoyment from it, simply because every time you re-architect an app, it introduces a load of risks of outage, bugs etc. Also, it's expensive, and ESO seem to make their main money from the Crown Store, so they will be loathe to risk that revenue stream for a load of architectural changes that they cannot monetise, until such time as they have to because its flakiness is losing lots of customers.

MaxwellC

@killmove
Honestly companies like ZOS follow the Law of Diminishing returns where as long as it's operational then it's fine. If this was creating a PR nightmare then ofc the company would probably put forth more resources to fix the issue but as this is a game I doubt it'll happen.

SirAndy

While the Hero engine may not be state of the art, the client side rendering has absolutely nothing to do with server side lag.

Nothing, nada, zilch, nix ...

FlyingSwan

MaxwellC wrote: »

@killmove
Honestly companies like ZOS follow the Law of Diminishing returns where as long as it's operational then it's fine. If this was creating a PR nightmare then ofc the company would probably put forth more resources to fix the issue but as this is a game I doubt it'll happen.

Yes, this ^^

If this was a serious line of business app that was causing loss of revenue then sure, they'd invest in it. But everything looks rosy in terms of server pop etc.

We do a lot of architecture work with the largest enterprises and they will stay on legacy stuff, that is way outside of support but that works, until the regulatory penalties for so doing outweigh the costs and risk of updating. When it comes to technology, many companies (often very well known global enterprises with hundreds of thousands of staff) adopt an 'if it ain't broke, don't fix it' mantra. It's myopic in my view, but IT budgets are controlled by bean counters obsessed with balancing the books on a yearly basis.

Tommy_The_Gun

I don't know is this somehow tied to Performance issues but It al started (and pretty much is going downhill since) when ZOS decided to drop support for directx 9.0 , 10 and 10,1.
I know that directx 9 is well... pretty archaic but I played the game back in 2014 using this mode and well performance was much better. Sure the game looks "slightly" better now, on directx 11 but the poor performance are simply not justified (when compared the old dx 9.0 rendering mode to present dx 11 mode the slightly better effects are simply not worth the performance loss).

Also I read somewhere (I can't remember where because it was some time ago) that DX 9.0 is a "native" Hero Engine rendering system. I can't tell if this is true but it might explain some performance issue we have today.

FlyingSwan

Tommy_The_Gun wrote: »

I don't know is this somehow tied to Performance issues but It al started (and pretty much is going downhill since) when ZOS decided to drop support for directx 9.0 , 10 and 10,1.
I know that directx 9 is well... pretty archaic but I played the game back in 2014 using this mode and well performance was much better. Sure the game looks "slightly" better now, on directx 11 but the poor performance are simply not justified (when compared the old dx 9.0 rendering mode to present dx 11 mode the slightly better effects are simply not worth the performance loss).

Also I read somewhere (I can't remember where because it was some time ago) that DX 9.0 is a "native" Hero Engine rendering system. I can't tell if this is true but it might explain some performance issue we have today.

Issue for ZOS would have been that DX9 dropped out of vendor extended support in April 2014. Dx9 was a Win XP component and MS operate a 10 year support roadmap, so at that point ZOS can no longer get support for it which impacts their ability to raise Premier calls with Microsoft if they have issues with code, get troubleshooting etc., also there are then no security patches which is bad for us and risks reputational damage to ZOS if there's any form of attack, so no developer will want to use anything outside of vendor support.

So it may be related to performance, but also would have been unavoidable. Technically however, each version of DX contains all that came before plus the new stuff. Usually, the improvements in the new release tend to improve performance of the older games anyway.

zyk

killmove wrote: »

Can we imagine ZOS doing a whole rework of the game by rebuilding and using a new engine (with all 2019 technologies) ?
they can make it so previous code (from actual engine) can be copied/pasted in the new one. that way they can keep all maps, models, etc. available in the game.

Don't know if we call this a remaster or a remake, I'm not a developper.

ESO's engine has never stopped being developed. There have been multiple major updates. I believe DX12 is on the roadmap.

Game development is actually in a very mature place and most engines are directly derived from engines that were used 10 years ago. Like most sophisticated software, they are too complex to constantly re-engineer from scratch so they are constantly updated.

IMO, the performance on the PVE-side of the game is great. The engine is actually very good at managing the huge number of art assets used in character customization.

It's mainly just Cyrodiil that's fundamentally broken.