How to build online community around CAD/PLM software?

October 13, 2014

community-engineers-CAD-PLM

There is one thing that seems make everyone interested and listen carefully these days – online communities. To build a successful community is a tricky thing. To make a money out of community is huge. Successful online communities can provide a lot of insight about how people are communicating, what is the value of community for different users and businesses. Communities are also a good reflection of business and software vendor ecosystem.

I’ve been discussing communities and marketplaces earlier on my blog. One of the most earliest discussion about that was Marketplace and Engineering software back in 2010. The interest of people to stay in the community is one of the most interesting factors to watch and learn. What will keep users to stay in the community and how to make it grow? Some of my earlier thoughts about that is here – PLM communities and let go threshold.

Online activity is a good indicator of community potential. I’ve been using blogging community as one of my own checkpoint about potential of users and customers to form a successful online community. Go online and try to find number of blogs about specific topic and their traffic. Do it for well-known MCAD packages (CATIA, Creo , Inventor, NX, SolidEdge, SolidWorks) as well as for PDM/PLM products. Google is of course not the best reflection of community size and activities. But it can give you some idea anyway.

CAD-PLM-communities-google-trends

Another interesting observation is related to how customers are sharing their knowledge. You can see very different behavior models. One of them is practically – share everything. This model is creating high viral content online. For this group it is not unusual to see how users are sharing their best practices and problems. Another group is conservative and closed. It is hard to get people to speak about this software. Most of information about it is curated either by software vendors and customers.

So, what are examples of successful in creating CAD/PLM communities? I can come with few examples. The list below is alphabetically sorted and I’m sure miss few communities (so, please don’t hesitate to suggest me additions to this group)

Aras. Back in 2007, Aras Corp. turned their Aras Innovator PLM product into so called “Enterprise Open Source”. This community demonstrated steady growth and I can see lots of information about Aras customers, open roadmap and open source development.

AutoCAD. The community of AutoCAD is big and reflect wide spread of this package in a very diverse set of verticals. I can see some association with community of users working with DWG. However, these users have the tendency to follow specific CAD packages developed around DWG formats.

GrabCAD. This is an interesting example of 1M+ engineers sharing 3D CAD models and related work online. GrabCAD website contains about 400K CAD models and this is probably one of the biggest in this kind.

SolidWorks. The community of SolidWorks is not a typical online (web) community. It was created around SolidWorks software back in before-web days. Nevertheless, this community demonstrated very high engagement level and shared lot of their work online in for the last 10+ years.

I guess “one million dollar question” is how to create a successful online engineering community. One of the myths of engineering world is that engineers are representing very anti-social group of people. The CreativeEngineer blog – A Few Common Myths About Engineers is actually bringing some opposite facts.

If you are thinking of being an engineer, you need to think of yourself as a leader, not a cubicle dwelling, anti-social, door mat. Here’s an except from Geoffrey C. Orsak, Dean of Engineering, SMU: In today’s reality, engineers are the new leadership class. Don’t believe me? Well, consider a recent survey of the S&P 500 CEOs by the global executive search firm SpencerStuart. Of these 500 key corporate leaders, nearly a quarter (23%) were educated as engineers and computer scientists.

So, is there an opportunity to create a larger community for engineers and users of CAD / PLM software? I guess the answer is yes. The shift towards online work is in a very beginning. My hunch existing engineering communities were just a beginning of future online working environment. The real community can be build around economic or social interest. In my view, the center of gravity will be moving from a specific CAD/PLM software towards companies and individuals. Internet is a new platform and community is a form to run business relationships on this platform.

What is my conclusion? Think about existing online communities: shoppers, open source software developers, education and many others. Try to make an analogy with manufacturing world. The economic and social interest is driving the most successful ones. I guess we just in the beginning of huge shift of engineering and manufacturing community towards online work. Economic interest will lead people to find new forms of business relationships and create new forms of CAD/PLM communities. Just my thoughts….

Best, Oleg


How PLM vendors will develop next security technologies

September 22, 2014

future-security-tech

The world is a scare place today. Every day we hear about different security problems – very large retail network informed me that their payment system was breached; large organization informes that customer databases with account information leaked, proprietary photos of celebrities available for general public, credit card numbers compromised. You can continue this list…

If you are responsible for IT organization, you should be scary to death. Think about engineering and manufacturing organization. How can you insure the protection of sensitive information assets? The time when a potential solution was just to cut a network cable into engineering working room is probably in the past. Even if you do that, mobile devices, global environment and need for supply chain access will simply put your effort in a trash can. EBNonline published a very interesting infographic called – The iPhone 6 Saga. It presents Apple’s global supply chain. I have no specific information about how Apple support security in their engineering, manufacturing and supply chains systems. What is clear – security is not optional there.

Security is well known topic in PLM community. Manufacturing companies are afraid of IP leaks for their most important intellectual assets. The question about how to protect data is on the top of the list when you evaluate PLM system. Introducing of cloud technologies and cloud PLM systems just raised the sense of urgency towards protecting highly sensible engineering and manufacturing assets.

Security of information in PLM system is multi-dimensional problem.. These days vendors are very vocal about their commitment to insure information is secured. It is easy to find this information available from vendors’ website and published white papers. A very simple Google search for “PLM security” returned a bunch of documents from Aras, Autodesk PLM360, Dassault Systems,  Siemens PLM, Oracle Agile PLM and few others. Take a look on links and draw your opinion.

It made me think, security management is not a topic that can be covered by one vendor. In modern manufacturing world, you hardly can find a company that using engineering, data management and manufacturing software from a single software vendor. The situation won’t be different in a future. It means manufacturing companies will have to insure security of sensitive IP. My hunch, software vendors will have to talk to develop future security technologies and systems together.

My attention caught THEINQUIRER article Google, Dropbox and others launch Simply Secure to ensure open security. Simply Secure is an organization that aims to work on how to solve the problem of security in an open source community.

A collection of firms and internet thinkers have gathered under a Simply Secure banner and promised to make security and privacy tools that protect and serve. Google and Dropbox are players in the Simply Secure support group and are joined by the Open Technology Fund. Together they all hope to address web users’ concerns about their content and connections. “Internet software links us to our friends, allows us to transact across oceans, and forms a digital space for culture and society. Because these technologies provide forums for sensitive discourse and expression, we are all concerned about their security and privacy.

The world of product lifecycle management is very competitive. Vendors are in a constant search how to differentiate their solution. However, security technology is a thing that can benefit a community of customers to adopt technologies and to work together. Modern enterprise software is quickly adopting best technological practices and tools coming from open source and web world. Security best practice can be next place where adopting of some industry and open source standards can be beneficial.

What is my conclusion? Security is a big deal. By developing future security technologies and adopting some well known best practices, PLM vendors can raise the bar for how they can protect customers from security breaches. The world around us is highly connected these days. To solve security problems is not an optional thing. It impacts every company in modern global manufacturing world.

Best, Oleg


What cloud PLM cannot do for you?

September 19, 2014

no-cloud-magic-plm

It has been already few years since I started to discuss cloud PLM opportunities on my blog. I found one of my early blogs about PLM and cloud – PLM and cloud: hold the promise?

So, what changed since then? Actually, quite a lot… We’ve seen massive adoption of cloud and mobile by businesses in many domains. PLM cloud adoption is growing too. Cloud is on the roadmap of all PLM vendors. It is really a question of "how to implement cloud?" rather than a question of "do we need to support cloud"? We also seen few very interesting examples of cloud applications in CAD/PLM space. I want just to mention few of them – Autodesk design tool Autodesk Fusion360, Dassault SolidWorks Mechanical Conceptual, Autodesk PLM360. Siemens PLM made their TeamCenter PLM available from IaaS infrastructure. Aras announced cloud strategy and introduced cloud product available via partnership with Infor ERP – Infor PLM Innovator. Cloud PLM pioneers, Arena Solutions, introduced several new cloud tools (BOM control and Quality management). Last but not least, GrabCAD, an open community of mechanical engineers released cloud PDM tool – GrabCAD Workbench. Earlier this week, GrabCAD was acquired by 3D printing company Stratasys. According to TechCrunch, article the deal was around $100M. I’m sure missed few products and companies…

Here are things that I discussed back in 2010 – cost of the solution, delivery model, global access, faster implementation, scaling. We learned a lot of about PLM and cloud for the last four years. Today, I want to make a reality check for list of things I discussed before in lights of what cloud PLM can or cannot do.

1- Cost

Cloud PLM made a mental switch in everything we knew about PLM before. According to Engineering.com article, cloud affected negatively on-premise PLM market. Cloud PLM created expectations for alternative pricing models and pushed all vendors to think how to turn PLM into service offering. Today, you can buy cloud PLM subscription with no upfront cost and hardware investment, which is a very good thing. However, I don’t think, total cost of ownership is different if you will calculate it on the period of 5 years. I’d love to see and learn more about that and love if you can share your comments.

2- Deployment, scale and IT

One of the best thing delivered by cloud PLM is related to deployment and IT cost. You can buy and deploy it instantly – almost similar to how you can open a new Gmail account. As a customer, you don’t need to worry about servers, setup cost, ordering hardware. You don’t need to negotiate with IT installation time. However, you cannot eliminate IT completely, especially if you are large company. For most of situations, you will have to discuss and make an alignment with IT about issues related to security and information access.

3- Faster Implementation

So, you can buy cloud PLM without upfront cost, you can deploy it overnight. What about PLM implementation? Implementation is an interesting thing. I’d like to speak about two aspects of implementations – 1/Configuration and customization; 2/ Implementations of business processes.

Four years ago, many companies were concerned about capability of web/cloud applications to deliver the level of flexibility, customization and configurations similar to on-premise PLM deployments. It is true, for most of situations, you cannot hack your cloud PLM with simple SQL script. However, I think, the flexibility of cloud PLM tools today is similar to on-premise PLM systems. However, flexibility of cloud PLM tools cannot provide real advantages compared to on-premise tools. Thanks for virtualization and modern collaboration technologies you can run your implementation remotely also for on-premise PLM systems.

Implementation of business processes is an interesting aspect of PLM implementation. In practice it means to define data structures and business processes. Cloud PLM won’t provide any advantages here. It is all about people, processes and organizational changes. So, the ugly truth is that cloud PLM won’t reduce your need of implementation services. In case of on-premise PLM, implementation will be done on site and collaborate with IT – installing, configuring and debugging customized software. In case of cloud PLM, you will need to work with cloud PLM vendor or hosting provider.

What is my conclusion? Cloud computing changed a lot in our life. PLM on the cloud can do many things differently. With much lower upfront cost and simple deployment, it opens PLM doors for many companies that never thought can buy and implement PLM systems before. However, when it comes to implementation and services, cloud PLM won’t do much different from on-premise PLM systems. You still need to implement it. It will require business process planning and implementation cost. Just my thoughts…

Best, Oleg


ERP vendors are ready to clash using PLM weapons

September 18, 2014

plm-erp-competition

The information about Aras PLM OEM deal with Infor caught my attention yesterday evening. It looks like a big deal for Aras Corp. Aras PLM is well known by their innovative Enterprise Open Source model. If you are not familiar with Aras, check their website and blog. For the last few years Aras demonstrated solid capabilities to deliver full scale PLM solution. You probably remember my old blog – Aras lines up against Windchill, ENOVIA and TeamCenter 3 years ago. I think Infor-Aras partnership is a step in the same direction. More information about Infor-Aras partnership is here. Product offering will be re-branded as Infor PLM Innovator. The following passage explains how does it fit Infor PLM strategy:

Infor PLM Innovator – powered by aras – provides a full-featured, highly scalable, flexible and secure PLM solution built on industry best practices that easily adapts to your company’s changing business practices. Seamlessly integrated with Infor ERP, Infor PLM Innovator goes beyond the capabilities of standalone PLM products, to unite your entire product lifecycle for a single view of the truth, which provides actionable information from design and manufacturing to purchasing, quality, the supply chain and beyond.

What is specially interesting, Infor is going to offer Aras PLM as a cloud-service. Consilia Vector research put an interesting perspective on an opportunity related to Aras-Infor deal.

Word of mouth and an open source sales program has taken Aras quite far; this deal with Infor will be like moving from cars to airplanes. It also gives Infor what it craves, a customer checklist that more completely matches item for item with Oracle and SAP. Oracle bought Agile PLM several years ago, and SAP has assembled a PLM system from a variety of small purchases and in-house projects.

I found comments about SAP and Oracle quite remarkable. It looks like all three top ERP vendors recognized the value and importance of Product Lifecycle Management. I guess, manufacturing and enterprise companies are challenged how to support innovation with constraints of global product development and manufacturing. Therefore, adding PLM functionality to their product suites is an important imperative for all of them.

The notion of cloud in Aras-Infor deal is interesting. It is going to challenge cloud ERP providers too. If you noticed, leading cloud ERP provider Netsuite made a strategic partnership alliance with Autodesk about integration between Netsuite ERP and Autodesk PLM 360.

So, what about top three CAD-PLM providers – Dassault, Siemens and PTC? I think, strategic focus on these vendors is outside of ERP-PLM bundles. Dassault is focusing on 3DExperience. PTC is developing strategy for IoT. Siemens is broadening their TeamCenter PLM platform and applications. All these vendors have some sort of partnership agreements with ERP vendors as well as love-hate relationships of selling PLM solutions to same manufacturing companies.

What is my conclusion? All ERP vendors are well armed with PLM weapon to provide a comprehensive product development and manufacturing solutions. It confirms growing adoption and interest of manufacturing companies of all sizes in PLM. It is time to grab some popcorn and get ready for PLM-ERP competitive performance. Just my thoughts…

Best, Oleg


The end of debates about out-of-the-box PLM?

September 8, 2014

plm-out-of-the-box

PLM implementation discussions are usually brings lots of controversy. Vendors, analysts, advisers, service companies, customers are all involved into implementations. It brings different and, sometimes, conflicting interests. In my view, one of the most debated topic in PLM implementations is related to so called ability to implement “PLM Out-of-the-box”. I’m not sure who first used that term. I think, it came out of PLM vendor marketing trying to demonstrate how easy and quick PLM implementation can be done. However, since then, the debates about “PLM Out of the box” had never ended.

Two other related topics are customization and configuration. For long period of time, I didn’t differentiate much between these two terms. However, modern enterprise software lexicon (and PLM vendors are in a full compliance with that) will define “configuration” (opposite to customization) a process that doesn’t require to write a software code for PLM implementation, but only use some elements of PLM user interface to configure a system. It probably turns all PLM implementation into “customization”, since writing programming scripts (using VB scripts or JS) is a widely used practice during all PLM configurations.

But let me get back to OOTB topic. I covered PLM OOTB few times in my blog. You can navigate your browser and read PLM Out-of-the-Box: Misleading or Focusing? published almost four years ago. From my latest posts, I can recommend you to take a read of the following article – Why My PLM won’t work for you? My attention was caught by an article that looks like trying to end all debates about PLM OOTB.

Aras Corp. published an interview with Dr. Martin Eigner who recently joined Aras’ board of advisors. In a very short published interview Dr. Eigner dots the i’s and crosses the t’s in the debates about out-of-the-box PLM and customization. Here is a main passage I captured. It has a strong Aras marketing flavor, but to quote it is important to bring a full message:

Dr. Martin Eigner: The kernel PLM functions are very similar from all competing PDM / PLM solution providers and from functionality it’s not a big criteria to differentiate each other. The user interface, performance and customization is important. Customization is very important because I do not believe even for small customers that you can buy PLM solution out of the box. That is a dream. You have to customize it. The real differentiator of existing PDM systems is the amount of money and capacity to customize a PLM solution. So I think usability, performance, upgrade capability and how easy it is to customize and maintain the customized solution are the most important points. They have the strongest impact on the total cost of ownership. I think in all these topics [performance, usability, upgrades, and customizations] Aras is leading. There are independent tests which show the system’s performance. We did internal tests at my university and found Aras to be the easiest to customize and upgrade. That is a big difference to the competitors. Customization is the most important aspect of PLM. Out-of-the-box works for no one.

I’m not in full agreement with Dr. Eigner about the fact you have to customize every PLM implementation. However, there is one point, which I think it is very important and I liked how Dr. Eigner emphasized that. It is related to the ability to maintain a customized PLM solution. This is one of the key differentiators of something I call a sustainable PLM platform. Customized legacy PLM is data management titanic in many implementations. Companies have to spend resources to maintain the solution, which is in most situation cannot support latest version of PLM platform provided by vendor.

What is my conclusion? Sustainable PLM platform. This is should be an important element of every PLM strategy these days. Modern business environment is very dynamic. Customer are looking for an agile way to implement business solution, adopt it to a new requirements as well as to maintain existing configuration. In my view, the concept of OOTB PLM should be revised with modern open architecture approach, which can simplify configuration, customization and sustainability of existing solutions. Just my thoughts…

Best, Oleg


Why unbundle 3D is hard for PLM vendors?

August 26, 2014

3d-cad-unbundle-plm-1

Unbundling is an interesting trend in many industries these days. It is relatively new marketing and business activity that helps to create new business offering, packages and product configurations. In many situations "unbundling" is a disruptive factors in many industries. Here is how it explained in Wikipedia article:

Unbundling is a neologism to describe how the ubiquity of mobile devices, Internet connectivity, consumer web technologies, social media and information access[1] in the 21st century are affecting older institutions (education, broadcasting, newspapers, games, shopping, etc.) by "break[ing] up the packages they once offered, providing particular parts of them at ascale and cost unmatchable by the old order."[2] Unbundling has been called "the great disruptor".[3] "Unbundling" most basically means simply the "process of breaking apart something into smaller parts."[4] In the context of mergers and acquisitions, unbundling refers to the "process of taking over a large company with several different lines of business, and then, while retaining the core business, selling off the subsidiaries to help fund the takeover."[5]

Enterprise software is well known by existing large "bundled" application suites. For long period of time, vendors developed large set of packaged applications. On the other side, customers’ demand was to achieve high level of vertical integration between product lines and product families. Last year, I explored some perspective on the future of unbundling in enterprise software and PLM. One of the drivers behind future "unbundling" is related to interests of customers to get better optimized software environment, focus on specific groups of users and driving faster ROI and fast implementations.

My attention caught my Aras blog post – If all you have is Teamcenter Everything Becomes a 3D CAD Problem. The article speaks exactly about the problem of bundles in engineering software. It discusses different needs of users in an organization. The split Aras introduced by Aras goes between people that need to get on 3D CAD software and rest of organization. Here is the passage, which explains that.

The 3D CAD vendors have created very complex file configuration management problems. Independent of how you manage your enterprise product lifecycle, you have to worry about breaking the configuration integrity of these fragile 3D CAD systems. Given the unique complexity of the 3D CAD problem, do you really expect that a single enterprise tool will be able to manage the entire product information data set and processes? Or is it better to manage CAD with the PDM system provided by the CAD vendor, and use a more suitable enterprise system to manage the majority of the product information and processes? Thousands of end users managing the true majority of product information and use cases have been asked to wait decades while exotic 3D CAD centric PLM systems are deployed to the specification and requirements of the few design engineers. But what is the missed opportunity cost to the business?

I can see Aras’ marketing and business message for "unbundling". As non-CAD PLM vendor, Aras is looking how to disrupt integrated suites provided by PLM vendors such as Siemens PLM and maybe others. At the same time, for customers looking how to solve a specific set of problems outside of engineering organization, to deliver such unbundled solution can be an interesting and efficient strategy.

There are lot of questions that customers will raise as soon as vendors like Aras will unbundle specific 3D CAD functionality from broader scope of process management. To achieve both vertical integration and granularity in platform and tools is very hard and this is a weak point in Aras strategy compared to integrated PLM suites. Few weeks ago, I debated that topic with Chad Jackson of Lifecycle Insight. Read about debates here – CAD: Engineering bundles vs. granular apps. More of my ideas and thoughts about the same topic is here – PLM: Tools, Bundles and Platforms.

What is my conclusion? To unbundle complex engineering applications suites as PLM is not easy. Vertical interesting is very important and it will be hard to give up them. Flexibility and agility are on the top priority lists for IT managers when it comes to management of application and resources these days. It looks like an interesting topic to put on the list for PLM vendors and software architects these days. Unbundling was very disruptive in many domains. Will PLM domain can be disrupted by unbundling into platforms and granular apps. Will 3D CAD become the first tool to unbundle from PLM? It is a good question to ask. Just my thoughts…

Best, Oleg


PLM upgrades, release cycles and legacy software

August 20, 2014

legacy-software

Do you know what is legacy software? Earlier today, Marc Lind of Aras Corp. challenged me by his twitter status about companies complaining about legacy PLM systems and upgrading. Here is the original passage from twitter here and here.

"a lot of people complains about legacy PLM and a lot of companies that have legacy PLM are throwing in the towel and switching these days".

marc-lind-legacy-plm-tweet

The part of statement about "legacy software" is really interesting. Last week, I wasn’t able to update a game on my son’s iPad. After few minutes, I discovered that Apple is not supporting the original iPad hardware manufactured 4 years ago. Does it mean iOS software run on that iPad is a legacy? Good question. At the same time, what about properly functioning ERP software that company runs already for the last 10 years without any plans to upgrade? Is that a legacy software?

Wikipedia gives me the following definition of legacy system:

In computing a legacy system is an old method, technology, computer system, or application program,"of, relating to, or being a previous or outdated computer system."[1] A more recent definition says that "a legacy system is any corporate computer system that isn’t Internet-dependent."[2]… The first use of the term legacy to describe computer systems probably occurred in the 1970s. By the 1980s it was commonly used to refer to existing computer systems to distinguish them from the design and implementation of new systems. Legacy was often heard during a conversion process, for example, when moving data from the legacy system to a new database.

Software upgrades is an important topic in engineering and manufacturing. Very often, systems can be in use very long time because of product lifecycle and the need to maintain existing data. It happens a lot in defense, aero and some other "regulated" industries. Also, because of significant investment, the ROI from upgrade can be questionable, which leads companies to keep existing outdated systems in operation. I’ve been posted about problems of PLM customization and upgrades before – How to eliminate PLM customization problems and Cloud PLM and future of upgrades.

PLM vendors are aware about the issue of upgrades and difficulties of software migrations . For long time, industry recognized it as something unavoidable. However, in today’s dynamic business environment, the issue of software upgrades cannot be ignored. Customers demanding flexible and agile software that can be deployed and updated fast. At the same time, changes of business models towards services and subscriptions pushed the problem of upgrades back to vendors.

Earlier this year, my attention was caught by CIMdata publication – Aras Innovator: Redefining Customization & Upgrades. Aras enterprise open source model is predominantly subscription oriented. Which provides lots of incentives for Aras engineers to solve the issue of upgrades and new versions deployment. Here is the passage from the article confirming that:

For several years, the Aras Corporation (Aras) has included no-cost version-to-version upgrades in their enterprise subscriptions, independent of how the solution has been customized and implemented. This is a rather bold guarantee given the historic challenges the industry has experienced with upgrading highly customized PLM deployments. With more than 300 upgrades behind it, CIMdata felt it appropriate to find out how Aras’ guarantee was playing out, and discovered that there was much more to the story than just a contractual guarantee. Fundamentally, Aras Innovator is engineered to be highly configurable—even customizable—without resulting in expensive and complex version-to-version upgrades and re-implementations.

One of PLM software leaders, Siemens PLM is also thinking about What is the best release cycle. The article speaks about SolidEdge release cycle.

A few years ago we moved from an irregular release cycle for Solid Edge, maybe 9 months in one cycle to 15 months in the next, to a regular cycle of annual releases (of course there are also maintenance packs delivered in the interim). I believe our customers much prefer this, they can plan ahead knowing that there will be a significant Solid Edge release available to them in August each year.

At the same time, the article confirms that CAD/PLM vendors are looking how to solve the problem of upgrades. As I mentioned earlier, cloud software model is one of the most promising technical ways to solve the issue of upgrades. It is true, but can be tricky in case both desktop and cloud software are involved. Here is the passage from the same Siemens PLM blog:

Working in the PLM area we try really hard to provide our customers with a good upgrade experience. PLM software is itself dependent on both the operating system and database software, and it has to work with specific releases of CAD software (sometimes with more than one CAD solution for our multi-CAD customers) and with office software as well! Moving PLM software to the cloud could potentially take some of the upgrade issues away from the end user, but PLM software does not work in isolation from your data files, or your other software and systems so I believe there is much work still to be done before the cloud really impacts the upgrade situation for real-world customers.

What is my conclusion? From customer perspective, the best option is to make release cycle completely transparent. In my view, this is really high bar for PLM vendors. Customer data migration, customization and sometimes absence of backward compatibility make release transparency questionable. However, since industry moves towards cloud software and service business model the demand for agile release management and absence of upgrades will be growing. So, my hunch, in the future we will not see "legacy software" anymore. New type of enterprise software will manage upgrades and migrations without customers paying attention. Sound like a dream? I don’t think so. For most of web and consumer software it is a reality already today. Just my thoughts…

Best, Oleg


Follow

Get every new post delivered to your Inbox.

Join 252 other followers