All Things Pivotal Episode #4 – Pivotal CF 1.3 – What's New?

October 28, 2014 Simon Elisha

featured-pivotal-podcastWhen investing in a Platform-as-a-Service, you want it to evolve quickly, providing new and useful capabilities on a regular basis. Pivotal CF is built using agile software development methodologies, meaning that customers get regular releases of new capability prioritized by customer demand.

Pivotal CF supports the on-line upgrade of versions using the powerful BOSH toolset, meaning that customers don’t just hear about new features–they can use them straight away!

One of the many benefits of PaaS is having the needs of multiple customers drive features that get wrapped into the platform–in effect, you get the benefits of the best thinking around software platforms and development delivered to you as software.

This week, we take a tour through some of the new capabilities to be found in Pivotal CF 1.3. Lots of handy goodies including High Availability improvements, security improvements, new language support and more.

PLAY EPISODE #4

 

RESOURCES:

Transcript

Simon:
Hello, everybody. Welcome back to the All Things Pivotal podcast. Fantastic to have you back. My name is Simon Elisha, CTO and Senior Manager of Field Engineering here in beautiful Australia and New Zealand. Both places at the same time but I do get across the ditch from time to time.

Awesome to have you back for episode number four. This one is going to be all about Pivotal CF 1.3 What’s New? A big release took place just recently and I thought I’d take it through step by step so we can cover what’s changed.

Now, Pivotal Cloud Foundry is now available version 1.3. This is the fourth release since we launched the products back into general availability in November 2013. What you are seeing here is a very fast cadence and a very reliable cadence of upgrades to the product to make sure that it is keeping pace with the best practice, current thinking and future thinking around platform as a service.

A raft of different capabilities had been included in this release, I’m going to take you through step by step of some of those components to show you what they mean and to show you what they can do for your organization.

One of the first changes is that the PCF now supports multiple availability zones. This means that you as a Cloud operator can correct many availability zones in the Pivotal CF operations manager and install Pivotal Cloud Foundry across clusters and resource pools. What this means from an operational perspective is should you experience any infrastructure issues in an availability zone, the other availability zones will take the workload and the applications keep on running.

We’re grabbing this handle, the workload capacities handle and it just keeps on going. Now, this is very important because things do happen, the unexpected happens. We’ll just have maintenance that you want to perform. The ability to support multiple availability zones gives you immense flexibility and immense durability in terms of how you drive your applications.

Another new change is support from multiple networks. In previous versions, the Pivotal Cloud Foundry components were deployed to a single network and the infrastructure administrators had to use isolation techniques using firewalls and writing rules, etc., to separate a management lab from the execution lab.

Now, you have the option to isolate Pivotal Cloud Foundry products on different networks. This is an important feature for those people who are really concerned with security and adds an extra larger protection beyond application containerization and encryption. If an applicational framework is breached, it can be isolated from corporate data and infrastructure so the breach is contained.

This gives you that good network segregation. A lot of the organizations who were comfortable with in terms of dividing up workloads in their environment. Again, fully supported and easy to do.

Another change is the ability to correct security groups. Security groups are fine grained policy-based wide list of rules which can be used by operators to set regulated application access to certain networks. This is for app-ban traffic. This can be targeted to spaces, so that you can provide the same policies around all applications in that space.

It allows you to set the protocol you’re using. It could be TCP, UDP, ICMP. The ports and the destination, be that particular destination or a CIDR range. What this means is you can control that app-ban communication piece. Particular applications need to speak to other applications on particular protocols. They can do it and you control it. Data company of course is in controlled through the ingress [routes 00:03:56] which manage across HTTP and HTTPS. They’re already controlled.

We now provide on demand access to enterprise to do. What this means is that Pivotal CF administrates can now define multiple service plans and each of the on demand service plans can correct different sized Pivotal HD, Hadoop Distribution, clusters comprised with different software components.

This means that you can build access for your developers into that data like component. You can correct the data piece. This is really useful in a Dev test environment whether if I may want to spin up a particular Hadoop cluster to test some data. Interactional. How the particular application and consumes data from the bigger corpus of data in the environment.

We also now have a support to interrupt [brightly 00:04:42] supported versions of AMC’s Isilon products. This can be configured to run Hadoop’s HDFS software component. What this means in this case is that the HDFS component is not co-located with the computer software, rather all data is written to install it on Isilon.

What this means is you can now manage your data separately from the particular nodes upon which you’re storing your information. It also allows you to do a whole lot of core capability of the Isilon support, I might go into that now. It makes it very easy for developers to dynamically point their application at the organizations production Pivotal HD environment as well. It can reach out to that external services to and still be running in a more defined fashionable recreated environment all the time and connects us that information very, very effectively.

It also means from a Dev test perspective, you can spin up environments very quickly, use them and manage their entire life circle. Service management within Cloud Foundry allows you to create services, access those services securely. Then, delete those services afterwards. Anyone who works in a development environment knows that that’s a life cycle that tends to take place very, very rapidly.

Some other changes have happened to the developer console. The Dev console is one of the two main interfaces for developers to use in Pivotal CF and is in addition to the Cloud Foundry CLI, the Command Line Interface. The Dev console is all about presenting information that’s contextual and is updated into real time.

In version 1.2 of the product, went a long way to improving that experience for developers and version 1.3 continues in this with two key improvements, the app dashboard and the usage reports. The app dashboard has been entirely redesigned to show you everything you need related to an app.

If that’s health, its scale, its status, its events. The [reps 00:06:35] that are being used, the service findings that it has and more, all on a single page. It provides you with in line and interactive controls, so you can scale your app’s instant count, the memory limit, the disk limit. You can start, stop or restart your app straight away. It gives you that really nice interactive experience.

It also gives you the real time user feedback. You’re seeing changes to scale or a state of your application immediately reflected in all elements of the app dashboard. You’re not playing the refresh game, you’re just seeing it in real time. There’s a live updating datas; new events and status changes continue to show up in a developer console whether it was initiated from the console or from the CLI or elsewhere, you’ll still see all the events coming back into that live data.

There’s also a new event viewer that shows you the ten most recent application events and it also allows you to do better organization and interaction with the bound service, instances and routes as well as additional information about your apps, build pack and stack and more. It’s giving you more insight into what’s going on in the environment.

There’s also a new usage report that shows the operator server [house 00:07:41]. People who are running this platform as a service. How much memory is used by each app? The data is derived from the app usage events in elastic run time. Things like stats, stops, scaling events and it’s rendered in gigabyte hours.

This is a major memory allocated overtime. We grouped these apps together in a parent space and all spaces in an organization sharing on the same page, so it’s easy to compare data. You can break down these data at the old layer or the space layer and the individual app layer which means that you can quickly see, ‘Is there an anomaly here or something is changing?’ Is it all balanced correctly.

There are detailed app drill downs as well. Instead of seeing a whole lot of events that have all cramped together, we can roll up all of these events into a particular time durations to make it much easier to visualize.

Of course, we have current and historical data. The default view is to show you the current calendar month. You can see what’s happening? What’s current? You can also see the previous two months of data as well and if you want to export that data, you can export all the data viewable on the page as a CSV file which means you can use it for other things as well.

That’s on the operational side. Of course, one of the benefits of Pivotal Cloud Foundry is the ability to create and bind to services. Some of the services we have are things like [inaudible 00:09:03] Pivotal, Pivotal HD which is Hadoop, RabbitMQ which is a message box, Reddit’s which is a key value case store. [Reaxys 00:09:11] which is a history compatible object store. [inaudible 00:09:14] which is a very well-known non-sequel database.

These services are integrated with the operations manager which is the interface upon which you manage your Cloud Foundry implementation. It allows that full life cycle management. This is quick through provisioning, console data logging, debugging, in flight updates and scaling. You can now also manage access to what’s called the marketplace.

These are services and plans that have made available to all the organizations or particular organizations within your group, so that you can allow developers to get instant access to a variety of these different services and applications which allows them to do testing, development and get that hands-on experience.

Using the service program, applications that developers push the Cloud Foundry can bind to these services automatically which reduces the cycle time because you get rid of that typical Cloud drive sort of having to figure out how I deploy the services, getting the security, getting the passwords, getting the networks setup, getting the resource management. All that goes away.

We just use this very clean service interface to consume the service as when we needed. Super it’s exciting for developers to be able to do that. Other capability of Cloud Foundry is the ability to deploy what we called build packs and the build packs allow you to deploy different language types on your platform and they understand what that language entails. What are the dependencies? How my build pack components are required? It’s a very powerful shortcut, if you like to getting deployed very, very quickly.

In version 1.3 we had that commercially supported build packs for Java, Ruby and Node.js, three very popular development languages.

We’re now adding support from PHP Python and Go Lang as well. Something I’m saying in the development’s sphere is this moves to a more polyglot approach to development and this ties into market’s services as well, but it’s really around developers using the right language for the task at hand, the one that gets the job done more efficiently.

The deploying different services running using different languages. Some are deploying now with Go Lang, others are sticking with Java because they believe it’s exactly what they need. Others have a preference for Node.js because they preferred the [inventing 00:11:24] model and etc. You can choose the one you like but be deploying in a consistent and coherent fashion with all the other components taken care of for you.

The logging, the scaling, the deploy process, etc., across development test and productions gets taken care of for you using the Cloud Foundry approach. It really is just a push it and adjust works approach. These commercially supported build packs now are laid to do that with a wider variety of language types.

Speaking of language types, the earth does not just run on English although it is certainly I think the most prevalent language on the internet today, although that might change in the future. We do want to ensure that we help customers from a variety of different countries with a variety primary languages.

In addition to English, the command line now supports four new languages. We now support French, Mandarin, Portuguese and Spanish. There’s also internationalization support for additional languages in the future through a plug-ability, capability that’s being built. Now, we’ll be able to see more languages supported in the future. You can use that command line in the language that most suits you, which is a great thing for people.

A great release, a lot in there. As you can see the velocity is not slowing down. The innovation is not slowing down. We’ll continue to do that and part of this is a really nice cycle where we look at how developers are using their platform and help them use it better. The things that can be taken care of for them, the features that are beneficial for them get rolled in very quickly.

These updates can be applied immediately to the platform and of course, because it’s a subscription-based licensing model, you get the updates automatically. You don’t have to worry about paying for new version, etc. You just get it. Have a look, have a play, some pretty exciting stuff there. Until then, I look forward to speaking to you again in another podcast and keep on building.

About the Author

Simon Elisha is CTO & Senior Manager of Field Engineering for Australia & New Zealand at Pivotal. With over 24 years industry experience in everything from Mainframes to the latest Cloud architectures - Simon brings a refreshing and insightful view of the business value of IT. Passionate about technology, he is a pragmatist who looks for the best solution to the task at hand. He has held roles at EDS, PricewaterhouseCoopers, VERITAS Software, Hitachi Data Systems, Cisco Systems and Amazon Web Services.

More Content by Simon Elisha
Previous
When in Doubt, Leave It Out
When in Doubt, Leave It Out

I’ve rarely regretted cutting a feature, whether working with clients at Pivotal Labs, as an in-house produ...

Next
Challenges With Randomness In Multi-tenant Linux Container Platforms
Challenges With Randomness In Multi-tenant Linux Container Platforms

Using Linux Containers for tenant isolation rather than OS isolation from VMs or non-virtualized hosts expo...

Enter curious. Exit smarter.

Register Now