Kalista IO is bringing artificial intelligence to storage. We created a revolutionary intelligent storage system reshaping how users experience, deploy, and manage storage



Photo of Lightwaves from creativity103.com, used under CC BY

Storage.
Simplified.

Phalanx storage system is self-managing, self-optimizing, and self-healing. It proactively identifies and prevents issues before they happen. And intelligently manages data, resources, and IOs for optimal performance. Phalanx makes storage effortlessly simple — so you have less to do and more to accomplish.


Ready. Set. Done.

Let Phalanx handle the ordinary. So you can be the extraordinary. From automatic indexing to self-optimizing data placement — Phalanx minimizes complicated setup and manual work. So you can focus on what really matters.


One giant leap.
For all IOs.

Intelligence is everywhere. From data path to control path — Phalanx brings the predictive and analytical powers of AI/ML to every operation. Read and write IOs are intelligently prioritized to reduce overall latency and contention. And management commands are optimized for maximum reliability and efficiency.




"The measure of intelligence is the ability to change." — Albert Einstein



All systems go.

Phalanx is proactive. So you don't have to be reactive. No one can stop failures but we can manage them. Phalanx tracks device and system level metrics to predict and manage availability and performance issues before they become issues.



Less is more.

When you don't have to do everything, you can do anything. Phalanx can be easily deployed with a single command. And works with all storage devices in any environment. It constantly learns and adapts to your access patterns to keep data secure and performance optimized.



Your IO. Your rules.

Versatility is built right in. Phalanx's AI/ML algorithms are always learning and improving — like your own personal team of dedicated storage experts. This makes Phalanx infinitely adaptable to new use cases and changing workloads. A perfect match to the dynamic software-defined world.

Keep your data sec•••.

Protection starts everywhere. Every access and change are monitored for deviations to protect the security and integrity of your data. It is easy to always have the latest and most secure version of Phalanx. When updates are available, you can opt-in to have them automatically installed when Phalanx is not in use.

One for all, all in one.

With Phalanx's unified object, block, and file interface — you now have a choice in data access format.

Work directly with cloud-native applications using S3 API.

Access files using traditional POSIX and NAS interfaces.

Address blocks through SCSI commands.

Easy to deploy.
Simple to operate.
Runs everywhere.


Simple. Transparent. Designed for users. Deploying Phalanx is as easy as downloading and running a virtual machine, a container image, or plugging in a hardware appliance. It just works.

Minimal dependency, maximum flexibility. Phalanx has minimal kernel dependency and requires no application changes. This means thinking less about what you need to do, more about what you want to do.

Deploy anywhere. Run everywhere. Deploy and operate Phalanx using existing orchestration and provisioning frameworks. It is designed to fit within your current workflows and environments. And seamlessly supports the latest storage device technologies such as SMR, HAMR, and multi-actuator.


Every type of data.
Any type of device.
All in one place.


Host Managed SMR support. An industry first. Seamlessly deploy industry leading next gen storage technology.





Massive, ultra-dense storage. Up to 26TB per device.

Reduced TCO. Say hello to power efficiency and a lower cost per capacity.

Consistent and predictable performance at every scale. Say goodbye to long tail latencies.

Cut latency at the tail

Device friendly and intelligent. Phalanx is designed to enable storage devices to perform at their best. Device awareness and semantic intelligence are built into every IO — enabling user experiences and performances that simply aren't found on any other storage system.

A picture is worth 600,000 IO.


Legacy stack with CMR


Phalanx with HM-SMR

Here is what happens when we modify a file 600,000 times.

The first IO latency histogram is from legacy software and a conventional HDD1. The large variability of response times (long tail) leads to unpredictable and inconsistent performance at scale.

Compare with the second histogram, from Phalanx and a Host Managed SMR HDD2 that shows a transformative difference — with faster and more consistent response times.



Up to

4.8x

lower latency3

Percentile latencies (usec)

Kalista Phalanx
with
Ultrastar HC620

Legacy stack
with
Ultrastar HC530

99%

16,924

28,468

99.95%

26,211

97,371

99.99%

41,736

202,227



There is magic in every IO. Each request is intelligently prioritized and streamlined to reduce contention and eliminate hot spots.

Phalanx be quick. Phalanx be nimble.

Our focus on performance and scalability means delivering consistent and predictable performance at every scale. So you can store and access data faster than ever.



Feature summary



Scales performance with capacity

Each storage device added to the system increases both performance and capacity.

Next gen ready

Ready for the latest in storage technologies. Phalanx natively supports Shingled Magnetic Recording (SMR), Heat-Assisted Magnetic Recording (HAMR), and variable capacity devices.

Seamless scaling to multicloud

Out of capacity? No problem! Burst & scale from on-prem to AWS, GCP, Azure, and more.

Maximum performance

No hot write areas and minimum device IO contention. Phalanx drives maximum performance out of each storage device.

Scalable and compatible

Virtualized and modular architecture. Self-contained storage VM/container fits within existing virtualization and orchestration frameworks.

Just works

Turnkey operation. No need to recompile or change existing applications. There is no kernel module to install. We've minimized the friction/disruption to deployment.

Technical information



Architecture

Designed to scale.
Engineered to perform.

Learn more >

Performance

Create without constraint.
Performance by design.

Learn more >

  1. Testing conducted by Kalista IO in July 2020 using XFS file system with Linux kernel 5.4.0-42-generic, and Intel® Core™ i7-4771 CPU 3.50GHz with 16GiB DDR3 Synchronous 2400 MHz RAM, and Western Digital Ultrastar DC HC530 CMR drive connected through SATA 3.2, 6.0 Gb/s interface. Write bench created a single 1GB file and executed 600,000 write commands each overwriting the first 64KB region of the file to capture latency values.
  2. Testing conducted by Kalista IO in July 2020 using preproduction Olympus (Phalanx) software with Linux kernel 5.4.0-42-generic, and Intel® Core™ i7-4771 CPU 3.50GHz with 16GiB DDR3 Synchronous 2400 MHz RAM, and Western Digital Ultrastar DC HC620 host managed SMR drives connected through SATA 3.2, 6.0 Gb/s interface. Write bench created a single 1GB file and executed 600,000 write commands each overwriting the first 64KB region of the file capture latency values.
  3. At 99.99th percentile latency. Testing conducted by Kalista IO in July 2020 using preproduction Olympus (Phalanx) software and XFS file system with Linux kernel 5.4.0-42-generic, and Intel® Core™ i7-4771 CPU 3.50GHz with 16GiB DDR3 Synchronous 2400 MHz RAM, and Western Digital Ultrastar DC HC620 host managed SMR and Ultrastar DC HC530 CMR drives connected through SATA 3.2, 6.0 Gb/s interface. Write bench created a single 1GB file and executed 600,000 write commands each overwriting the first 64KB region of the file capture latency values.
  4. Photo of Lightwaves from http://creativity103.com, CC BY 3.0 Used under CC BY / scaled from original.