ScaleXtreme Reins in Clouds with Orchestration 

Cloud management software provider ScaleXtreme is puffing up its third major release of its cloud management tools, adding a workflow and orchestration engine that can reach up into the applications running on various public and private clouds and make them behave.

ScaleXtreme was founded in August 2010 by Nand Mulchandani and Balaji Srinivasa, both of whom have long careers in systems software and want to take on the incumbents in systems management in this cloudy era. Those incumbents are IBM, BMC, CA, and Hewlett-Packard, who all have very sophisticated management frameworks and large installed bases. Mulchandani contends that their tools are too expensive and, at six to nine months for a typical installation, take far too long to get up and running.

"The old big four in management are not going to be the new big four," Mulchandani tells EnterpriseTech.

ScaleXtreme wants to be one of the new big four, and it is fleshing out its product, which is itself a cloudy service, to add more functionality that enterprise customers want.

Srinivasa, who is CTO at ScaleXtreme, designed the Operations Manager tool from Bladelogic, which BMC paid $800 million to acquire in March 2008 to give it virtualization management tools. Before that, he was a software engineer at IBM, Computer Associates (now known simply as CA), and Platinum Technology. After leaving Bladelogic and concurrent with the impending cloud boom, Srinivasa started coding what would eventually become the ScaleXtreme tool, which has had a few names in the past few years (Xpert was the first one) and which is now just called ScaleXtreme like the company. Mulchandani was CEO at OpenDNS before joining Srinivasa as ScaleXtreme's CEO, and before that he was senior director of product management and marketing at VMware in charge of its vShield hypervisor security product. ScaleXtreme has raised $13.5 million in two rounds of funding from Accel Partners and Ignition Partners to get off the ground.

The ScaleXtreme cloud management tool is itself running on a homegrown cloud that is hosted in two datacenters in San Francisco and Chicago (their locations are not disclosed) with backup capacity for disaster recovery set up on the Amazon Web Services and Rackspace Hosting public clouds. The ScaleXtreme management tool has a mix of SQL and NoSQL databases to store configuration and performance information for both physical and virtual servers. The back-end code is written in Java and the agents that are plunked onto virtual or physical instances are written in C to give them a tiny footprint and fast performance. The user interface for ScaleXtreme is written in a mix of JavaScript and JQuery and is designed to let system admins manage infrastructure from a PC, tablet, or smartphone of their choosing.

The neat trick about ScaleXtreme is that you don't have to open up ports on the firewall for your infrastructure to be linked through its agents to the ScaleXtreme cloud. All you need is an encrypted HTTP link and ScaleXtreme's agents can talk out of your virtual or physical machines and be controlled by them. Exactly how this is done without a proper two-way link is, in fact, a trade secret.

Over 100,000 virtual machines are under management by the ScaleXtreme service, which has thousands of customers. (Mulchandani will not be more precise about the customer count.) That is a lot of VMs all aggregated together, of course, but the management job is not computationally intensive so it doesn't take a zillion servers to support the ScaleXtreme app. How many it does take, Mulchandani did not say.

Conceptually, here is what the ScaleXtreme tool looks like:

With the 3.0 release of the tool, ScaleXtreme is getting a self-service portal and a service catalog so lines of business and other end users can help themselves – if they are authorized, of course – to infrastructure and software to run on it. The updated code now includes application blueprinting as well to manage the provisioning of multiple-tier applications and has features to do autoscaling in a consistent way across private and public clouds alike. The ScaleXtreme tool already did provisioning, access control, and budgeting of capacity on Amazon Web Services as well as any cloud based on OpenStack or VMware's vCloud/ESXi combination. These features came out in the initial release of ScaleXtreme from 18 months ago. With the next major update, ScaleXtreme was able to do patch management of Windows and Linux instances running on virtual or physical machines.

The ScaleXtreme service comes in a number of different editions. The Basic Edition is free and does monitoring of operating systems running on VMs or bare metal, showing CPU, disk, memory, and I/O usage as well as the processes that are running. It gathers up performance metrics and allows system administrators to set thresholds for triggers and alerts. Mulchandani says that there are tens of thousands of customers using this freebie tool, and like other fermium models, the intent is to give some of the tool away to encourage customers to pay for additional features.

ScaleXtreme has an additional server and application monitoring module that adds on top of this, which costs $2 per physical or virtual machine per month. This is the part of the stack that includes REST-based APIs, which allows admins to programmatically control ScaleXtreme and in turn their infrastructure with nary a mouse click. This module already supports the Apache Web server, the MySQL and SQL Server databases and the Redis and MongoDB data stores out of the box, along with a list of other popular infrastructure software.

If you want the patch management module, then in costs another $2 per machine per month.

For the full-tile-boogie version, which is called Advanced Cloud Management and which has the workflow and orchestration tools that debut with the 3.0 release included, the cost ranges from $10 to $15 per machine per month, depending on which modules you have underneath activated.

This full-on ScaleXtreme tool can hook into AWS, Microsoft Windows Azure, Google Compute Engine, Rackspace Cloud Servers, HP Cloud Services, Bluelock, and Verizon Terremark Enterprise Cloud public clouds. It offers full integration with the Chef configuration management tool from Opscode and the Github code repository.

The latest ScaleXtreme software has been under beta testing with unnamed Fortune 500 clients to manage their Azure, Terremark, and Azure cloud instances.